Target identification method based on laser radar, laser radar system and storage medium

Abstract

The embodiment of the invention discloses a target identification method based on a laser radar, a laser radar system and a storage medium. The method comprises the steps that a first laser radar and a second laser radar are used for scanning a target object, and a first point cloud collected by the first laser radar and a second point cloud collected by the second laser radar are obtained; obtaining a first point cloud translation amount and a second point cloud translation amount according to the plurality of detection distances and the parameters corresponding to the plurality of laser radars; performing point cloud fusion on the first point cloud and the second point cloud according to the first point cloud translation amount and the second point cloud translation amount to obtain a fused point cloud; and performing target identification according to the fused point cloud. According to the method, the point cloud translation amount is calculated based on the detection distance and the parameters of the laser radar, the point cloud is corrected according to the point cloud translation amount, and the point cloud fusion precision and the detection precision of the laser radar are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of laser radar, in particular to a target recognition method based on laser radar, a laser radar system and a storage medium. BACKGROUND

[0002] As an active ranging device, laser radar has been widely used in automatic driving, industrial mapping, robots and intelligent transportation and other fields. The continuous expansion of application fields also puts forward higher requirements for the detection performance of laser radar. In order to realize long-distance measurement of a large field of view, the existing technology usually performs multi-field-of-view splicing based on multiple laser radars or multiple transceiver modules. Among them, limited by the system architecture of the laser and the receiver device, the optical system of the existing laser radar usually adopts an off-axis scheme to realize the transmission and reception of laser beams. Due to the position difference between the transmitting lens and the receiving lens of such off-axis optical system, pixel shift will occur in the point cloud image actually received by the receiver device, and the closer the detection distance, the greater the pixel shift. When such pixel shift occurs in the overlapping field of view of multiple laser radars, it will cause blind spots in the point cloud corresponding to the overlapping field of view and affect the point cloud fusion effect, thereby affecting the detection accuracy of the laser radar. SUMMARY

[0003] To improve the detection accuracy of the laser radar, the present application discloses a target recognition method based on laser radar, a laser radar system and a storage medium.

[0004] In a first aspect, the present application discloses a target recognition method based on laser radar, which comprises:

[0005] scanning a target object using a first laser radar and a second laser radar to obtain a first point cloud collected by the first laser radar and a second point cloud collected by the second laser radar;

[0006] obtaining a first point cloud translation amount according to a first detection distance and parameters of the first laser radar, wherein the first detection distance is the distance between the target object and the first laser radar;

[0007] obtaining a second point cloud translation amount according to a second detection distance and parameters of the second laser radar, wherein the second detection distance is the distance between the target object and the second laser radar;

[0008] performing point cloud fusion on the first point cloud and the second point cloud according to the first point cloud translation amount and the second point cloud translation amount to obtain a fused point cloud;

[0009] performing target recognition according to the fused point cloud.

[0010] In some embodiments, the laser radar comprises a transmitting assembly and a receiving assembly, the receiving assembly comprises a receiving lens and a receiving chip, the transmitting assembly of the first laser radar is located at a first side of the receiving assembly of the first laser radar, the second laser radar is located at a second side of the receiving assembly of the first laser radar, and the parameters of the first laser radar comprise a baseline distance, an image focal length and a pixel pitch, wherein the baseline distance is the distance between the transmitting assembly of the first laser radar and the receiving assembly of the first laser radar, the image focal length is the focal length of the receiving lens of the first laser radar, and the pixel pitch is the size of one pixel on the receiving chip of the first laser radar.

[0011] In some embodiments, the laser radar comprises a first transmitting assembly, a second transmitting assembly and a receiving assembly, the receiving assembly comprises a receiving lens and a receiving chip, the receiving assembly of the first laser radar is located between the first transmitting assembly of the first laser radar and the second transmitting assembly of the first laser radar, the receiving assembly of the first laser radar is located at a first side of the second transmitting assembly of the first laser radar, the second laser radar is located at a second side of the second transmitting assembly of the first laser radar, and the parameters of the first laser radar comprise a baseline distance, an image focal length and a pixel pitch, wherein the baseline distance is the distance between the first transmitting assembly of the first laser radar and the receiving assembly of the first laser radar, the image focal length is the focal length of the receiving lens of the first laser radar, and the pixel pitch is the size of one pixel on the receiving chip of the first laser radar.

[0012] In some embodiments, the first point cloud translation is obtained according to the first detection distance and the parameters of the first laser radar, comprising: an image offset angle is obtained according to the first detection distance and the baseline distance of the first laser radar; an image offset distance is obtained according to the image offset angle and the image focal length of the first laser radar; and the first point cloud translation is obtained according to the image offset distance and the pixel pitch of the first laser radar.

[0013] In some embodiments, the first point cloud and the second point cloud are fused according to the first point cloud translation and the second point cloud translation to obtain a fused point cloud, comprising: the first point cloud is translated according to the first point cloud translation to obtain a translated first point cloud; the second point cloud is translated according to the second point cloud translation to obtain a translated second point cloud; and the translated first point cloud and the translated second point cloud are fused to obtain the fused point cloud.

[0014] In a second aspect, the present application discloses a laser radar system, comprising a first laser radar and a second laser radar;

[0015] The laser radar comprises a transmitting assembly and a receiving assembly, the transmitting assembly comprises a transmitting lens and a laser, and the receiving assembly comprises a receiving lens and a receiving chip.

[0016] The transmitting assembly of the first laser radar is located on a first side of the receiving assembly of the first laser radar, and the second laser radar is located on a second side of the receiving assembly of the first laser radar.

[0017] An included angle between an optical axis of the transmitting lens of the first laser radar and an optical axis of the transmitting lens of the second laser radar is greater than zero.

[0018] In some embodiments, the laser radar comprises a first transmitting assembly, a second transmitting assembly and the receiving assembly; wherein the receiving assembly of the laser radar is located between the first transmitting assembly of the laser radar and the second transmitting assembly of the laser radar.

[0019] In some embodiments, the receiving assembly of the first laser radar is located on a first side of the second transmitting assembly of the first laser radar, and the second laser radar is located on a second side of the second transmitting assembly of the first laser radar.

[0020] In some embodiments, a first preset angle is equal to a second preset angle; wherein the first preset angle is an included angle between an optical axis of the transmitting lens of the first transmitting assembly of the first laser radar and an optical axis of the receiving lens of the receiving assembly of the first laser radar; and the second preset angle is an included angle between an optical axis of the transmitting lens of the second transmitting assembly of the first laser radar and an optical axis of the receiving lens of the receiving assembly of the first laser radar.

[0021] In a third aspect, the embodiments of the present application provide a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the target identification method of the laser radar point cloud in the above embodiments.

[0022] The present application discloses a target identification method based on a laser radar, a laser radar system and a storage medium. In the point cloud fusion process for two laser radars, the point cloud translation amount corresponding to the center field of view in each point cloud is calculated according to the relative position relationship of the multiple transmitting assemblies and the multiple receiving assemblies and the parameters of each laser radar, so as to effectively solve the center field of view blind area problem after splicing, improve the point cloud fusion effect, and effectively improve the detection accuracy of the laser radar. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used by the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description only constitute some embodiments of the present application.

[0024] Figure 1 is a schematic diagram of an off-axis optical system disclosed by the embodiments of the present application;

[0025] Figure 2 is a schematic diagram of a laser radar system disclosed by the embodiments of the present application;

[0026] Figure 3 is a schematic diagram of a point cloud fusion method corresponding to the laser radar system of Figure 2 ;

[0027] Figure 4 is a schematic diagram of a laser radar system disclosed by the embodiments of the present application;

[0028] Figure 5 is a schematic diagram of a point cloud fusion method corresponding to the laser radar system of Figure 4 ;

[0029] Figure 6 is a schematic diagram of a laser radar system disclosed by the embodiments of the present application;

[0030] Figure 7 is a schematic diagram of a point cloud fusion method corresponding to the laser radar system of Figure 6 ;

[0031] Figure 8 is a schematic diagram of a laser radar system disclosed by the embodiments of the present application;

[0032] Figure 9 is a schematic diagram of a point cloud fusion method corresponding to the laser radar system of Figure 8 ;

[0033] Figure 10 is a schematic diagram of an off-axis optical system disclosed by the embodiments of the present application;

[0034] Figure 11 is a schematic diagram of a laser radar system disclosed by the embodiments of the present application;

[0035] Figure 12 is a schematic diagram of a point cloud fusion method corresponding to the laser radar system of Figure 11 .

[0036] Legend of reference signs: 100, laser radar; 101, transmitting assembly; 101a, transmitting lens; 101b, laser; 102, receiving assembly; 102a, receiving lens; 102b, receiving chip. DETAILED DESCRIPTION

[0037] For the purpose of making the purpose, technical solutions and advantages of the present application clearer, the embodiments of the present application will be further described in detail below with reference to the drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of structures consistent with some aspects of the present application as detailed in the appended claims.

[0038] As an active ranging device, the ranging performance of a laser radar is affected by various factors, such as power, stray light, size of receiving aperture, focal length, and field of view angle, etc. If the chip size of a laser radar is relatively fixed, the ranging performance and angular resolution will decrease if the field of view angle needs to be increased; if the ranging performance and angular resolution need to be improved, the field of view angle will decrease. In order to simultaneously satisfy the ranging performance, angular resolution, and field of view angle, the prior art usually performs point cloud splicing based on at least two laser radars or multiple transceiver modules to achieve long-distance measurement of a large field of view. Among them, the transceiver module of the laser radar usually adopts an off-axis optical system, and the difference in installation position between the transmitting lens and the receiving lens in such an off-axis optical system will cause pixel shift problems of the point cloud. Moreover, the closer the detection distance, the greater the pixel shift amount. When such pixel shift occurs in the overlapping field of view of multiple laser radars, a blind area will appear in the point cloud corresponding to the overlapping field of view, and the point cloud fusion effect will be affected, thereby affecting the detection accuracy of the laser radar.

[0039] In one example, as Figure 1As shown, the transceiver module of the laser radar includes a transmitting assembly 101 and a receiving assembly 102, wherein the transmitting assembly 101 includes a transmitting lens 101a and a laser 101b, and the receiving assembly 102 includes a receiving lens 102a and a receiving chip 102b. The optical axis of the transmitting lens 101a is parallel to the optical axis of the receiving lens 102a. The laser 101b is located at the focal plane of the transmitting lens 101a, and the receiving chip 102b is located at the focal plane of the receiving lens 102a. The laser 101b includes one or more laser emitting units, which are symmetrically distributed on both sides of the optical axis of the transmitting lens 101a. The receiving chip 102b includes one or more photosensitive units, and each photosensitive unit includes one or more pixels. The photosensitive units are symmetrically distributed on both sides of the optical axis of the receiving lens 102a. Due to the difference in installation position between the transmitting lens 101a and the receiving lens 101b, there is a difference in spatial direction between the scanning light beam emitted by the transmitting assembly 101 and the echo light beam actually received by the receiving assembly 102, so that the transmitting horizontal field of view and the receiving horizontal field of view cannot be completely matched, and further, a pixel shift is generated between the ideal imaging position and the actual imaging position on the imaging surface of the receiving chip 102b. In one example, the image side shift angle θ = arctan (L1 / D), and the image side shift distance H = f*θ between the ideal imaging position and the actual imaging position on the imaging surface of the receiving chip 102b, wherein the baseline distance L1 is the distance between the optical axis of the transmitting lens 101a and the optical axis of the receiving lens 102a or the distance between the optical center of the transmitting lens 101a and the optical center of the receiving lens 102a, D is the detection distance (the distance between the target object and the laser radar), and f is the focal length of the receiving lens 102a. Correspondingly, the pixel shift P = H / S between the ideal imaging position and the actual imaging position on the imaging surface of the receiving chip 102b, wherein the pixel pitch S is the size of a single pixel on the receiving chip. In the case where L1 is fixed, the closer the detection distance D is, the larger the image side shift angle θ is, and correspondingly, the larger the image side shift distance H and the pixel shift P are, which causes part of the pixels on the receiving chip 102b to be unable to receive the echo light beam. The point cloud area corresponding to the part of the pixels is a field of view blind area, which reduces the horizontal field of view angle or the vertical field of view angle of the laser radar and affects the point cloud accuracy after multiple point clouds are spliced.

[0040] In one embodiment, the present application discloses a laser radar system, which includes a first laser radar and a second laser radar. As shown in the drawings, Figure 2 As shown, one laser radar 100 includes a transmitting assembly 101 and a receiving assembly 102. The transmitting assembly of the first laser radar is located at a first side of the receiving assembly of the first laser radar, and the second laser radar is located at a second side of the receiving assembly of the first laser radar. The first laser radar is located at a first side of the receiving assembly of the second laser radar, and the transmitting assembly of the second laser radar is located at a second side of the receiving assembly of the second laser radar. In one example, Figure 2The angle between the optical axes of the emitting lenses of the first and second lidars shown is greater than zero. This arrangement causes the laser emission directions of the first and second lidars to be offset, which can further expand the field of view after stitching and compensate for the reduction in field of view caused by pixel offset.

[0041] In one example, such as Figure 2 As shown, considering the mismatch between the receiving and transmitting horizontal fields of view caused by the off-axis system, the actual horizontal field of view angle of the first lidar is α, the actual horizontal field of view angle of the second lidar is γ, and the horizontal field of view angle corresponding to the overlapping field of view of the first and second lidars is β. The field of view regions of the first lidar other than the overlapping field of view are the edge fields of view of the first lidar, and the field of view regions of the second lidar other than the overlapping field of view are the edge fields of view of the second lidar. Since the receiving component of the first lidar is closer to the second lidar than the transmitting component, and the receiving component of the second lidar is closer to the first lidar than the transmitting component, pixel offset issues will occur in the point cloud regions corresponding to the overlapping field of view for both the first and second point clouds acquired by the first and second lidars. This will affect the point cloud fusion effect and reduce the detection accuracy of the lidar system.

[0042] To improve Figure 2 To improve the detection accuracy of the lidar system shown, this application discloses a target recognition method based on lidar. Before point cloud stitching or fusion, this method first corrects each point cloud instance with pixel offset issues in the overlapping field of view region, thereby improving the point cloud fusion accuracy. In one embodiment, the method includes:

[0043] S101. Use a first lidar and a second lidar to scan the target object and obtain a first point cloud and a second point cloud, wherein the first point cloud is collected by the first lidar and the second point cloud is collected by the second lidar.

[0044] S102. Based on the first detection distance and the parameters of the first lidar, obtain the first point cloud translation amount.

[0045] In one example, the parameters of the first lidar include a first baseline distance L1, a first image-side focal length f1, and a first pixel pitch S1. According to the first baseline distance L1 and a first detection distance D1, a first image-side offset angle θ1 = arctan(L1 / D1) is obtained; according to the first image-side offset angle θ1 and the first image-side focal length f1, a first image-side offset distance H1 = f1*θ1 is obtained. According to the first image-side offset distance H1 and the first pixel pitch S1, a first pixel offset P1 = H1 / S1 is obtained. The first detection distance D1 is the distance between the target object and the first lidar. The first baseline distance L1 is the distance between the emission component of the first lidar and the receiving component of the first lidar. The first image-side focal length f1 is the focal length of the receiving lens of the receiving component of the first lidar. The first pixel pitch S1 is the size of one pixel on the receiving chip of the receiving component of the first lidar.

[0046] Since the receiving component of the first lidar is closer to the second lidar than the emission component of the first lidar, the first point cloud collected by the first lidar will have a pixel offset problem in the part of the point cloud corresponding to the overlapping field of view. Therefore, the first point cloud translation amount is equal to the first pixel offset P1 calculated, wherein the first pixel offset P1 is greater than a first preset threshold value, and the first preset threshold value is 1, 2, or 3 pixels. The detection distance corresponding to the first preset threshold value is taken as a second preset threshold value. When the detection distance is greater than the second preset threshold value, the point cloud translation amount is small, and the pixel offset problem of the first point cloud is not considered. In one example, when the first preset threshold value is 1 pixel and the first detection distance is greater than the second preset threshold value, the first point cloud translation amount is less than one pixel, and no translation correction is performed on the first point cloud. In another example, when there are multiple target objects in the detection field of view of the first lidar system, the detection distance of the target object closest to the first lidar among the multiple target objects is taken as the first detection distance.

[0047] S103, obtaining a second point cloud translation amount according to the second detection distance and the parameters of the second lidar.

[0048] In one example, the parameters of the second lidar include a second baseline distance L2, a second image-side focal length f2, and a second pixel pitch S2. According to the second baseline distance L2 and the second detection distance D2, a second image-side offset angle θ2 = arctan(L2 / D2) is obtained. According to the second image-side offset angle θ2 and the second image-side focal length f2, a second image-side offset distance H2 = f2*θ2 is obtained. According to the second image-side offset distance H2 and the second pixel pitch S2, a second pixel offset P2 = H2 / S2 is obtained. The second baseline distance L2 is the distance between the transmitting assembly of the second lidar and the receiving assembly of the second lidar. The second image-side focal length f2 is the focal length of the receiving lens of the receiving assembly of the second lidar. The second pixel pitch S2 is the size of one pixel on the receiving chip of the receiving assembly of the second lidar.

[0049] Since the receiving assembly of the second lidar is closer to the first lidar than the transmitting assembly of the second lidar, the second point cloud collected by the second lidar will have a pixel offset problem in the part of the point cloud corresponding to the overlapping field of view. Therefore, the calculated second pixel offset P2 is used as the second point cloud translation, where the second pixel offset P2 is greater than the first preset threshold.

[0050] S104, point cloud fusion is performed on the first point cloud and the second point cloud according to the first point cloud translation and the second point cloud translation, to obtain a fused point cloud.

[0051] In one example, as shown in Figure 3 the first point cloud is translated according to the first point cloud translation to obtain a translated first point cloud; the second point cloud is translated according to the second point cloud translation to obtain a translated second point cloud; and the translated first point cloud and the translated second point cloud are fused to obtain a fused point cloud.

[0052] S105, target recognition is performed according to the fused point cloud.

[0053] In one embodiment, the present application discloses a lidar system as shown in Figure 4 the receiving assembly of the first lidar is located on the first side of the transmitting assembly of the first lidar, and the second lidar is located on the second side of the transmitting assembly of the first lidar. The transmitting assembly of the second lidar is located on the second side of the receiving assembly of the second lidar, and the first lidar is located on the first side of the receiving assembly of the second lidar. Compared with Figure 2The laser radar system shown, although the first point cloud collected by the first laser radar still has a pixel offset, but the pixel offset at this time appears in the part of the first point cloud corresponding to the edge field of view of the first laser radar, which has no overlapping part with the second point cloud, and does not affect the subsequent point cloud fusion effect, so the first point cloud translation amount is recorded as 0, and the second point cloud translation amount is still equal to the second pixel offset P2, so there is no need to correct the first point cloud.

[0054] In one embodiment, for Figure 4 The laser radar system shown, the application discloses a laser radar-based target recognition method, which comprises the following steps:

[0055] S201, scanning a target object using a first laser radar and a second laser radar to obtain a first point cloud and a second point cloud, wherein the first point cloud is collected by the first laser radar, and the second point cloud is collected by the second laser radar.

[0056] S202, obtaining a second point cloud translation amount according to the second detection distance and the parameters of the second laser radar.

[0057] S203, performing point cloud fusion on the first point cloud and the second point cloud according to the second point cloud translation amount to obtain a fused point cloud.

[0058] In one example, as Figure 5 shown, performing a translation transformation on the second point cloud according to the second point cloud translation amount to obtain a translation-transformed second point cloud; and performing point cloud fusion on the first point cloud and the translation-transformed second point cloud to obtain a fused point cloud.

[0059] S204, performing target recognition according to the fused point cloud.

[0060] In one embodiment, the application discloses a laser radar system as Figure 6 shown, wherein the emission assembly of the first laser radar is located on the first side of the receiving assembly of the first laser radar, and the second laser radar is located on the second side of the receiving assembly of the first laser radar. The first laser radar is located on the first side of the emission assembly of the second laser radar, and the receiving assembly of the second laser radar is located on the second side of the emission assembly of the second laser radar. Compared with Figure 2 the laser radar system shown, although the second point cloud collected by the second laser radar still has a pixel offset problem, but the pixel offset at this time appears in the part of the second point cloud corresponding to the edge field of view of the second laser radar, which has no overlapping part with the first point cloud, and does not affect the subsequent point cloud fusion effect, so the first point cloud translation amount is still equal to the first pixel offset P1, but the second point cloud translation amount is equal to 0, so there is no need to correct the second point cloud.

[0061] In one embodiment, forFigure 6 The present application discloses a target recognition method based on lidar for the lidar system shown, the method comprising:

[0062] S301. Use a first lidar and a second lidar to scan the target object and obtain a first point cloud and a second point cloud, wherein the first point cloud is collected by the first lidar and the second point cloud is collected by the second lidar.

[0063] S302. Based on the first detection distance and the parameters of the first lidar, obtain the first point cloud translation amount.

[0064] S303. Based on the point cloud translation amount of the first point cloud, the first point cloud and the second point cloud are fused to obtain the fused point cloud.

[0065] In one example, such as Figure 7 As shown, the first point cloud is translated based on the translation amount to obtain the translated first point cloud; the first and second point clouds are then fused to obtain the fused point cloud.

[0066] S304. Target recognition is performed based on the fused point cloud.

[0067] In one embodiment, this application discloses as follows Figure 8 The illustrated lidar system has a receiving component located on a first side of a transmitting component, and a second lidar located on a second side of the transmitting component. The first lidar is located on a first side of the transmitting component, and the receiving component is located on a second side of the transmitting component. Compared to... Figure 2 In the illustrated lidar system, the pixel shift of the first point cloud occurs within a portion of the first point cloud corresponding to the edge field of view of the first lidar, and this portion of the point cloud does not overlap with the second point cloud. Similarly, the pixel shift of the second point cloud occurs within a portion of the second point cloud corresponding to the edge field of view of the second lidar. Therefore, the translation amount of both the first and second point clouds is zero, and no translation correction is required for either. In one example, such as... Figure 9 As shown, the first point cloud and the second point cloud are stitched together to obtain a stitched point cloud; the stitched point cloud is then fused to obtain a fused point cloud; and the fused point cloud is used for target recognition.

[0068] In one embodiment, in Figure 1 Based on the off-axis optical system shown, this application discloses as follows: Figure 10The off-axis optical system shown includes a first emission component, a second emission component and a receiving component. The receiving component is located between the first emission component and the second emission component. In one example, a first preset angle is equal to a second preset angle, wherein the first preset angle is an included angle between an optical axis of an emission lens of the first emission component and an optical axis of a receiving lens of the receiving component; and the second preset angle is an included angle between an optical axis of an emission lens of the second emission component and the optical axis of the receiving lens of the receiving component. In one example, a laser of the first emission component is located on a first side of the optical axis of the emission lens of the first emission component, and a laser beam of the laser of the first emission component is emitted toward a second side of the optical axis of the emission lens of the first emission component. A laser of the second emission component is located on a second side of the optical axis of the emission lens of the second emission component, and a laser beam of the laser of the second emission component is emitted toward a first side of the optical axis of the emission lens of the second emission component. Compared with Figure 1 The off-axis optical system shown, Figure 10 The off-axis optical system shown matches the emission horizontal field of view of the two emission components with the receiving horizontal field of view of the receiving component, which can reduce the size of the light emitting surface of a single laser and improve the light emitting uniformity of the single laser.

[0069] In one embodiment, based on Figure 10 The off-axis optical system shown discloses a laser radar system, as shown in Figure 11 The receiving component of the first laser radar is located between the two emission components of the first laser radar. The receiving component of the first laser radar is located on a first side of the second emission component of the first laser radar, and the second laser radar is located on a second side of the second emission component of the first laser radar. The receiving component of the second laser radar is located between the two emission components of the second laser radar. The first laser radar is located on a first side of the first emission component of the second laser radar, and the receiving component of the second laser radar is located on a second side of the first emission component of the second laser radar. The included angle between the optical axis of the receiving lens of the first laser radar and the optical axis of the receiving lens of the second laser radar is greater than zero.

[0070] In one embodiment, for Figure 11 The laser radar system shown discloses a laser radar-based target identification method, which includes:

[0071] S401, using the first laser radar and the second laser radar to scan a target object to obtain a first point cloud and a second point cloud, wherein the first point cloud is collected by the first laser radar, and the second point cloud is collected by the second laser radar.

[0072] S402, obtaining a first point cloud translation amount according to a first detection distance and parameters of the first laser radar.

[0073] In one example, the parameters of the first lidar include a baseline distance, a first image focal length f1 and a first pixel pitch S1. The first point cloud has a pixel offset P1a in the part of the point cloud corresponding to the edge field of view of the first lidar, and the pixel offset P1a corresponds to a first distance L1a. The first point cloud has a pixel offset P1b in the part of the point cloud corresponding to the overlapping field of view, and the pixel offset P1b corresponds to a second distance L1b. Wherein the first distance L1a is the distance between the first emitting component of the first lidar and the receiving component of the first lidar, and the second distance L1b is the distance between the second emitting component of the first lidar and the receiving component of the first lidar. The baseline distance of the first lidar is equal to the first distance L1a, the first pixel offset is equal to the pixel offset P1a, and the first point cloud translation is equal to the first pixel offset.

[0074] S403, obtaining a second point cloud translation according to the second detection distance and the parameters of the second lidar.

[0075] In one example, the parameters of the second lidar include a baseline distance, a second image focal length f2 and a second pixel pitch S2. The second point cloud has a pixel offset P2a in the part of the point cloud corresponding to the edge field of view of the second lidar, and the pixel offset P2a corresponds to a third distance L2a. The second point cloud has a pixel offset P2b in the part of the point cloud corresponding to the overlapping field of view, and the pixel offset P2b corresponds to a fourth distance L2b. Wherein the third distance L2a is the distance between the first emitting component of the second lidar and the receiving component of the second lidar, and the fourth distance L2b is the distance between the second emitting component of the second lidar and the receiving component of the first lidar. The baseline distance of the second lidar is equal to the fourth distance L2b, the second pixel offset is equal to the pixel offset P2b, and the second point cloud translation is equal to the second pixel offset.

[0076] S404, performing point cloud fusion on the first point cloud and the second point cloud according to the first point cloud translation and the second point cloud translation, to obtain a fused point cloud.

[0077] In one example, as shown in Figure 12 , performing translation transformation on the first point cloud according to the first point cloud translation to obtain a translation-transformed first point cloud; performing translation transformation on the second point cloud according to the second point cloud translation to obtain a translation-transformed second point cloud; and performing point cloud fusion on the translation-transformed first point cloud and the translation-transformed second point cloud to obtain a fused point cloud.

[0078] S405, performing target recognition according to the fused point cloud.

[0079] In some embodiments, the laser 101b can be one of a fiber laser, a vertical-cavity surface-emitting laser (VCSEL), or an edge-emitting laser (EEL). The receiving chip 102b is a planar array receiving chip including a plurality of single photon avalanche diodes (SPADs). In one embodiment, the SPAD planar array receiving chip includes a plurality of columns of SPADs, and a pixel offset corresponding to each point cloud can be obtained according to a difference between a number of columns actually receiving light energy and a total number of columns of the plurality of columns of SPADs.

[0080] In some embodiments, the transmitting lens includes a beam expanding lens or prism, a first collimating lens, and the receiving lens includes a converging lens and a second collimating lens. The laser beam emitted by the laser passes through the first collimating lens to be subjected to divergence angle constraint, and then passes through the beam expanding lens or prism to be subjected to light deflection, so as to increase the scanning field angle in the horizontal direction or the vertical direction, thereby expanding the scanning range. The echo light beam reflected by the target object passes through the converging lens to be subjected to light deflection to reduce the beam divergence angle, and then passes through the second collimating lens to make the echo light beam converge on the light sensitive surface of the receiving chip.

[0081] In one embodiment, the present application provides a computer readable storage medium (or a non-transitory machine readable storage medium or a machine readable storage medium) having stored thereon executable code (or computer program or computer instruction code) which, when executed by a processor of an electronic device (or a server, etc.), causes the processor to perform part or all of steps S101-S105, part or all of steps S201-S204, part or all of steps S301-S304, and part or all of steps S401-S405.

[0082] In the description of the present application, it should be understood that, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description herein is for describing specific embodiments only and is not intended to be limiting of the present application. As used in the description herein, the term "and / or" includes all combinations of one or more of the associated listed items. The singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. The terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, elements, and / or components, but do not preclude the presence or addition of one or more other features, elements, components, and / or groups thereof. The terms "first" and "second" used in the embodiments of the present application are only for identification, and do not indicate other meanings such as a specific order or imply relative importance.

[0083] In the present application, unless otherwise clearly specified and limited, "on" or "under" of a first feature to a second feature can include that the first and second features are in direct contact, or that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, "on", "above" and "over" of a first feature to a second feature includes that the first feature is directly above and obliquely above the second feature, or only means that the first feature is higher than the second feature in horizontal height. "Under", "below" and "underneath" of a first feature to a second feature includes that the first feature is directly below and obliquely below the second feature, or only means that the first feature is lower than the second feature in horizontal height. The specific meanings of the above terms can be understood according to specific circumstances by those of ordinary skill in the art. "One or more embodiments" used herein do not refer to the same embodiment, but combine specific features, structures or properties in any suitable manner. The above is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A target recognition method based on lidar, characterized in that, The method includes: The target is scanned using a first lidar and a second lidar to obtain a first point cloud collected by the first lidar and a second point cloud collected by the second lidar. Based on the first detection distance and the parameters of the first lidar, the first point cloud translation is obtained, wherein the first detection distance is the distance between the target and the first lidar; Based on the second detection distance and the parameters of the second lidar, the second point cloud translation is obtained, wherein the second detection distance is the distance between the target and the second lidar; The first point cloud and the second point cloud are fused based on the translation amount of the first point cloud and the translation amount of the second point cloud to obtain the fused point cloud; Target identification is performed based on the fused point cloud.

2. The method according to claim 1, wherein the lidar includes a transmitting component and a receiving component, the receiving component includes a receiving lens and a receiving chip, the transmitting component of the first lidar is located on a first side of the receiving component of the first lidar, and the second lidar is located on a second side of the receiving component of the first lidar, characterized in that, The parameters of the first lidar include baseline distance, image-side focal length, and pixel pitch. The baseline distance is the distance between the transmitting component and the receiving component of the first lidar. The image-side focal length is the focal length of the receiving lens of the first lidar. The pixel pitch is the size of a pixel on the receiving chip of the first lidar.

3. The method according to claim 1, wherein the lidar comprises a first transmitting component, a second transmitting component, and a receiving component, the receiving component comprising a receiving lens and a receiving chip, the receiving component of the first lidar being located between the first transmitting component and the second transmitting component of the first lidar, the receiving component of the first lidar being located on a first side of the second transmitting component of the first lidar, and the second lidar being located on a second side of the second transmitting component of the first lidar, characterized in that, The parameters of the first lidar include baseline distance, image-side focal length, and pixel pitch. The baseline distance is the distance between the first transmitting component and the receiving component of the first lidar. The image-side focal length is the focal length of the receiving lens of the first lidar. The pixel pitch is the size of a pixel on the receiving chip of the first lidar.

4. The method according to claim 2 or 3, characterized in that, The step of obtaining the first point cloud translation amount based on the first detection distance and the parameters of the first lidar includes: The image offset angle is obtained based on the first detection distance and the baseline distance of the first lidar; The image-side offset distance is obtained based on the image-side offset angle and the image-side focal length of the first lidar. The first point cloud translation amount is obtained based on the image offset distance and the pixel spacing of the first lidar.

5. The method according to claim 1, characterized in that, The step of fusing the first point cloud and the second point cloud based on the first point cloud translation amount and the second point cloud translation amount to obtain the fused point cloud includes: The first point cloud is translated based on the translation amount of the first point cloud to obtain the translated first point cloud. The second point cloud is translated based on the translation amount of the second point cloud to obtain the translated second point cloud. The point clouds are fused based on the first point cloud after translation transformation and the second point cloud after translation transformation to obtain the fused point cloud.

6. A lidar system, characterized in that, The lidar system includes a first lidar and a second lidar, wherein the lidar includes a transmitting component and a receiving component, the transmitting component includes a transmitting lens and a laser, and the receiving component includes a receiving lens and a receiving chip; The transmitting component of the first lidar is located on the first side of the receiving component of the first lidar, and the second lidar is located on the second side of the receiving component of the first lidar. The angle between the optical axis of the emitting lens of the first lidar and the optical axis of the emitting lens of the second lidar is greater than zero.

7. The system according to claim 6, characterized in that, The lidar includes a first transmitting component, a second transmitting component, and the receiving component; The receiving component of the lidar is located between the first transmitting component and the second transmitting component of the lidar.

8. The system according to claim 7, characterized in that, The receiving component of the first lidar is located on the first side of the second transmitting component of the first lidar, and the second lidar is located on the second side of the second transmitting component of the first lidar.

9. The system according to claim 7, characterized in that, The first preset angle is equal to the second preset angle; Wherein, the first preset angle is the angle between the optical axis of the emitting lens of the first emitting component of the first lidar and the optical axis of the receiving lens of the first lidar. The second preset angle is the angle between the optical axis of the transmitting lens of the second transmitting component of the first lidar and the optical axis of the receiving lens of the first lidar.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, is used to implement the method as described in any one of claims 1 to 5.

Images