Laser radar angle resolution enhancement method and system

Abstract

The invention relates to the technical field of laser radars, and provides a laser radar angular resolution enhancement method, which comprises the following steps of: obtaining a spatial direction priority according to a vehicle course, road information and target information obtained by a vehicle external sensor so as to dynamically adjust laser pulse density and energy level; scanning data generated after N circles of continuous scanning serve as a composite frame, the starting angle of each circle is shifted, multiphase filtering and coherent superposition are conducted in a digital domain, and the effective angular resolution is obtained. The invention further discloses a laser radar device, the method and the device can adjust the pulse density and energy distribution in the ROI direction in real time according to the dynamic scene of the vehicle, the total pulse number is kept unchanged, and the total power consumption is not increased. The effective horizontal angle resolution is improved through inter-frame phase offset and multi-phase synthesis. A corner reflector and an IMU (Inertial Measurement Unit) are integrated in the laser radar, and phase and ranging errors are corrected in real time through periodic loopback calibration, so that the angle stability is kept.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of laser radar, and in particular to a laser radar angle resolution enhancement method and system. BACKGROUND

[0002] In the automatic driving technology system, the laser radar (LiDAR, Light Detection and Ranging) is one of the core sensors for realizing environment perception, which calculates the distance, position, shape and motion state of the target by emitting laser beams and receiving reflected signals, and provides high-reliability three-dimensional environment data for the automatic driving system. Its application runs through the whole process of automatic driving perception, positioning, decision-making and planning, and is one of the key technologies for solving how to clearly see the surrounding environment of automatic driving.

[0003] CN114202461B discloses a high-resolution processing method, device and computer storage medium. The high-resolution processing method comprises: acquiring a low-resolution point cloud; projecting the low-resolution point cloud into a first distance map according to the angle parameters of the laser radar, wherein the first distance map comprises a distance map of a coordinate channel and a distance map of a distance channel, the distance map of the coordinate channel comprises an x channel map, a y channel map and a z channel map, x, y and z represent three-dimensional information of the point cloud, the distance map of the distance channel is a range channel map, range represents distance, and the first distance map is input into a pre-trained neural network, wherein the neural network comprises a resolution network and a channel attention network, the distance map of the distance channel is processed by the resolution network to obtain a high-resolution distance map, and the distance map of the coordinate channel is processed by the channel attention network to obtain an attention distance map; the second distance map is obtained by adding the attention distance map and the high-resolution distance map; the second distance map is inversely projected to obtain a high-resolution point cloud; wherein, before the first distance map is input into the pre-trained neural network, the high-resolution processing method further comprises: acquiring the horizontal resolution of the first distance map; calculating the left padding number based on the horizontal resolution; calculating the right padding number based on the left padding number and the horizontal resolution; and performing cyclic padding on the first distance map based on the left padding number and the right padding number.

[0004] CN120314967A discloses a laser measurement method, a laser radar and an autonomous vehicle. The method comprises: generating a first laser beam and a second laser beam, wherein each of the first laser beam and the second laser beam is a frequency-modulated laser, has the same sweep frequency period and different wavelengths, and the frequency variation directions of the first laser beam and the second laser beam are opposite within the sweep frequency period; multiplexing the first laser beam and the second laser beam into a sweep frequency beam; splitting the sweep frequency beam into a signal beam and a local oscillator beam; emitting the signal beam; receiving a reflected beam generated by reflection of the signal beam after encountering an object; performing in-phase quadrature phase coherent demodulation on the local oscillator beam and the reflected beam. In order to obtain the scalar value of the beat frequency of the up-conversion stage and the scalar value of the beat frequency of the down-conversion stage between the local oscillator beam and the reflected beam; and detecting the beat frequency of the up-conversion stage and the beat frequency of the down-conversion stage between the local oscillator beam and the reflected beam to determine the speed of the object and / or the distance between the object and the laser radar.

[0005] The horizontal beam width of the existing laser radar is usually in the range of 0.1°-0.5°. In a long-distance detection scene, it is difficult to accurately capture the lateral detail features of the target, and it cannot meet the demand of high-precision resolution of the target at a long distance for automatic driving, which may cause the recognition accuracy of the long-distance obstacle to decrease. SUMMARY

[0006] Long-term practice shows that the traditional laser radar adopts a fixed scanning mode, lacks the ability to dynamically adjust the beam resources according to the real-time driving scene, and cannot optimize the beam allocation for the key monitoring directions of different scenes, such as the dense pedestrian area at the city intersection and the front vehicle lane-changing area on the highway, thereby causing the contradiction of insufficient sampling in the key direction, inability to obtain sufficient data to support target recognition, and excessive sampling in the remaining direction, causing waste of beam resources, and affecting the detection efficiency and practicability of the laser radar.

[0007] Therefore, the present application aims to provide a vehicle peripheral safety zone recognition method based on distributed ultrasonic waves, which comprises,

[0008] Step S1: calculating a spatial direction priority P according to the vehicle heading, road information and target information obtained by a vehicle external sensor; wherein the vehicle external sensor at least includes a camera and a millimeter wave radar;

[0009] Step S2: adjusting the laser pulse density and energy level in the single circle scanning of the laser radar according to the spatial direction priority;

[0010] Step S3: taking the scanning data generated after continuous scanning of N circles as a composite frame, and the starting angle offset is θ beamN, multi-phase filtering and coherent superposition are performed in the digital domain to obtain effective angular resolution θ eff , where N is a positive integer greater than 1, θ beam is the inherent beam width of the laser radar.

[0011] In an embodiment, in step S3, multi-phase filtering and coherent superposition processing are performed by including FFT or MUSIC super-resolution algorithm to obtain point cloud data.

[0012] In an embodiment, the point cloud data is subjected to quadratic interpolation to obtain effective angular resolution θ eff = θ beam / (2 N ).

[0013] In an embodiment, the vehicle heading includes a heading angle; and the road information includes road curvature.

[0014] In an embodiment, the spatial direction priority P of each spatial direction is calculated, where the spatial direction priority P is positively correlated with the time to collision TTC, the heading angle cosine, and the target category weight.

[0015] In an embodiment, in step S2, the pulse density ρ and / or laser energy E of the ROI direction are increased by a first preset proportion according to the spatial direction priority P value, where,

[0016] ∑ρ i ≤ρ total

[0017] ∑E i ≤E total

[0018] In a non-ROI direction, the pulse density ρ and / or laser energy E are reduced by a second preset proportion, so that the total pulse number ρ total and the total power consumption E total do not increase, where ρ i is the pulse density of the i-th direction, and E i is the laser energy of the i-th direction.

[0019] In an embodiment, an angle reflector and an inertial measurement unit are arranged in a housing of the laser radar, a loop calibration is triggered once per preset time, a phase error and a distance measurement error are calculated in real time by using the inertial measurement unit data and the echo data, and the phase error and the distance measurement error are written into a FPGA lookup table for phase correction of a next synthetic frame.

[0020] The application further discloses a laser radar device according to the laser radar angle resolution enhancement method.

[0021] A priority calculation unit is used for calculating a spatial direction priority P of each spatial direction by using vehicle heading, road information and target information obtained by a vehicle external sensor;

[0022] A pulse energy distribution unit is used for increasing a pulse density p and / or laser energy E of an ROI direction by a first preset proportion according to a value of the spatial direction priority P, and decreasing the pulse density p and / or the laser energy E of a non-ROI direction by a second preset proportion according to the value of the spatial direction priority P;

[0023] A phase offset multi-phase synthesis module is used for synthesizing scanning data generated after N continuous scans into a synthetic frame, and calculating an effective angle resolution;

[0024] A data interface module is used for defining and analyzing an Ethernet frame, wherein the Ethernet frame format comprises an Ethernet frame header, a self-defined header and M groups of 4 Byte data units, wherein M<=362;

[0025] An online self-calibration module is used for loop calibration of the laser radar and phase correction of a next synthetic frame, and comprises an angle reflector and an inertial measurement unit, and the angle reflector comprises a spiral optical fiber loop.

[0026] The application provides an electronic device, at least one processor; and

[0027] A memory in communication connection with the at least one processor; wherein,

[0028] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above-mentioned laser radar angle resolution enhancement method.

[0029] The application provides a machine readable storage medium, and the machine readable storage medium stores instructions for causing a machine to execute the above-mentioned laser radar angle resolution enhancement method.

[0030] The laser radar angle resolution enhancement method disclosed by the application, through steps S1-S3, obtains a spatial direction priority according to the vehicle heading, road information and target information obtained by the vehicle external sensor, dynamically adjusts the laser pulse density and energy level according to the spatial direction priority, and generates scanning data after continuous scanning of N circles as a composite frame, each circle has a starting angle offset, and multi-phase filtering and coherent superposition are performed in the digital domain to obtain an effective angle resolution. The application also discloses a laser radar device for the laser radar angle resolution enhancement method. The method and the laser radar device can adjust the pulse density and energy distribution of the ROI direction in real time according to the dynamic scene of the vehicle, keep the total pulse number unchanged, and do not increase the total power consumption. Without changing the laser and optical aperture, the effective horizontal angle resolution is improved through inter-frame phase offset and multi-phase synthesis. The angle reflector and IMU are integrated in the laser radar, the phase and ranging errors are corrected in real time through periodic loop calibration, and the 0.005° level angle stability is maintained.

[0031] Other features and advantages of the present application will be described in detail in the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0032] The accompanying drawings, which form a part of the present application, are included to provide a further understanding of the application, and are incorporated in and constitute a part of this specification. The illustrations are shown to explain the present application and are not intended to limit the present application. In the drawings:

[0033] Figure 1 A comparison diagram of the point cloud coverage area of the traditional uniform scanning and the point cloud coverage area of the laser radar angle resolution enhancement method of one embodiment of the application.

[0034] Figure 2 A comparison diagram of the traditional uniform scanning and the enhanced scanning of the laser radar angle resolution enhancement method of one embodiment of the application.

[0035] Figure 3 A schematic diagram of the Ethernet frame format in the laser radar angle resolution enhancement method of one embodiment of the application.

[0036] Figure 4 A block diagram of the laser radar system structure in the laser radar angle resolution enhancement method of one embodiment of the application.

[0037] Figure 5 A block diagram of the online self-calibration system in the laser radar angle resolution enhancement method of one embodiment of the application. DETAILED DESCRIPTION

[0038] The specific embodiments of the present application are described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely intended to illustrate and explain the present application, and are not intended to limit the present application.

[0039] In order for those skilled in the art to better understand the technical scheme of the present application, the technical scheme in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0040] It should be noted that the terms "first", "second", "third" and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not necessarily limit to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0041] In order to solve the technical problems existing in the prior art, that is, the traditional laser radar adopts a fixed scanning mode, lacks the ability to dynamically adjust the beam resource according to the real-time driving scene, and cannot optimize the beam allocation for different scene key monitoring directions, such as urban intersection pedestrian dense area and highway front vehicle lane-changing area, thereby appearing the contradiction of insufficient sampling in key directions, unable to obtain sufficient data to support target identification, and excessive sampling in the remaining directions, causing waste of beam resources, affecting the detection efficiency and practicability of the laser radar, etc. The present application provides a laser radar angle resolution enhancement method, as shown in Figures 1-5 The laser radar angle resolution enhancement method of an embodiment of the present application is shown in the schematic diagram,

[0042] Step S1, according to the vehicle heading, road information and target information obtained by the vehicle external sensor, the spatial direction priority P is calculated; wherein the vehicle external sensor at least includes a camera, a millimeter wave radar;

[0043] Step S2, within a single circle scanning of the laser radar, the laser pulse density and energy level are adjusted according to the spatial direction priority;

[0044] Step S3, the scanning data generated after N continuous scanning is taken as a composite frame, and the starting angle offset θ of each circlebeam / N, multi-phase filtering and coherent superposition are performed in the digital domain to obtain effective angular resolution θ eff wherein N is a positive integer greater than 1, and θ beam is the inherent beam width of the laser radar.

[0045] The laser radar angle resolution enhancement method disclosed in the application, through steps S1-S3, obtains a spatial direction priority according to the vehicle heading, road information and target information obtained by the vehicle external sensor, dynamically adjusts the laser pulse density and energy level according to the spatial direction priority, and generates scanning data after continuous scanning of N circles as a composite frame, with a starting angle offset every circle, multi-phase filtering and coherent superposition are performed in the digital domain to obtain effective angular resolution. This method can adjust the pulse density and energy distribution of the ROI direction in real time according to the dynamic scene of the vehicle, keep the total number of pulses unchanged, and not increase the total power consumption. Without changing the laser and optical aperture, through inter-frame phase offset and multi-phase synthesis, the effective horizontal angular resolution is improved. The angle reflector and IMU are integrated inside the laser radar, and the phase and ranging errors are corrected in real time through periodic loop calibration, so that the angle stability is maintained.

[0046] In order to better synchronize the acquisition of multi-dimensional key information and associate the calculation of the spatial direction priority, in step S1, the vehicle heading, road curvature, external sensor, including the target direction and spatial direction priority given by the camera / millimeter wave, are obtained in real time. Through the vehicle's own heading detection module, such as inertial measurement unit IMU, global navigation satellite system GNSS, etc., the driving heading data of the current vehicle is dynamically obtained, and the motion direction reference of the vehicle itself is determined. Secondly, combined with high-precision map data or vehicle chassis sensors, such as road curvature information feedback by the steering angle sensor, the path form characteristics of the current driving section are mastered, and the road environment basis for judging the potential key attention direction is provided. At the same time, the external sensor system composed of cameras and millimeter wave radars is connected, in which the camera is responsible for identifying the target, such as the two-dimensional image information of pedestrians, vehicles, traffic signs and matching their direction relative to the vehicle, and the millimeter wave radar accurately captures the three-dimensional spatial position and motion state of the target through ranging and speed measurement function, and further confirms the actual direction of the target. Finally, the vehicle heading based on the motion reference itself, the road curvature and the target direction perceived by the external sensor, i.e. the core attention object, are comprehensively sorted by priority in different spatial directions through a preset algorithm, and the spatial direction priority result directly used for subsequent laser radar regulation is generated. In order to better calculate, the preset algorithm includes weighted calculation based on risk level, target distance and road matching degree.

[0047] In order to improve the angular resolution and signal-to-noise ratio of the point cloud, so as to obtain higher resolution and higher precision point cloud data. In a more preferred embodiment of the present application, in step S3, the point cloud data is obtained by performing multi-phase filtering and coherent superposition processing including Fast Fourier Transform (FFT) or Multiple Signal Classification (MUSIC) super-resolution algorithm. The laser radar scans N times, and the starting angle of each circle is offset by θ beam / N in turn, forming N groups of original point cloud data with phase difference, and each group of data corresponds to a scanning circle with angular offset. Time synchronization is performed on the N groups of data, that is, the scanning time difference and the coordinate system are calibrated, that is, unified to the vehicle coordinate system, to ensure that the data can be superimposed in space, and finally the data in time and space is aligned. Coherent superposition is to perform phase calibration on the filtered N groups of target signals according to the known angular offset θ beam / N, to ensure that the signals of each circle are in phase in space, and phase compensation is realized. The compensated signals are coherently superimposed, and the signal-to-noise ratio weighted summation is preferably used to enhance the amplitude of the target signal, and the signal-to-noise ratio is improved by N times, while random noise is suppressed, and non-coherent noise is averaged and cancelled. The superimposed signal is converted back to the spatial domain, and the inverse FFT is preferably used for FFT processing, and three-dimensional point cloud is generated in combination with the laser radar ranging data. Through the algorithm, the physical limitations of the hardware are broken through, and the point cloud can more clearly distinguish adjacent targets at close range. Coherent superposition suppresses noise, making the point cloud of weak reflection targets, such as pedestrians and non-metallic obstacles, more stable.

[0048] In order to obtain point cloud with higher density and higher effective angular resolution. Virtual points calculated are inserted between the physical sampling intervals, so as to fill the information gap in the angle dimension. In a more preferred embodiment of the present application, the point cloud data is twice interpolated to obtain an effective angular resolution θ eff = θ beam / (2 N ). The core of the twice interpolation is to construct a quadratic function based on the angle and distance information of the adjacent three known points, to calculate the distance value at any angle position between two points, including Lagrange quadratic interpolation and Newton quadratic interpolation. If the distance obtained by interpolation is too different from the distance of the adjacent sample points, or exceeds the effective ranging range of the laser radar, it is determined as an invalid interpolation point and is removed, so as to avoid introducing false targets. All original points and interpolation points are reordered by angle to form a new high-density point cloud with smaller angle interval, and the effective angular resolution of the point cloud at this time is the target angle interval. The interpolated point cloud is locally smoothed to further suppress the possible micro fluctuations introduced in the interpolation process, and the continuity of the point cloud edge is ensured.

[0049] To avoid relying on a single sensor system, which is prone to failure in bad weather, no lane lines, etc. scene, the fusion of heading angle and road curvature can significantly improve adaptability. In a more preferred embodiment of the present application, the vehicle heading includes the heading angle; the road information includes the road curvature. The vehicle heading angle (Heading Angle) refers to the angle between the vehicle longitudinal axis and the true north direction or the relative angle with the road reference line, which is used to determine the vehicle's own driving direction. Based on the global navigation satellite system GNSS, through satellite signals such as GPS / Beidou, the absolute heading angle of the vehicle in the earth coordinate system is directly output.

[0050] The sensors such as lidar and camera can dynamically adjust resources to avoid redundant processing of low-risk areas and reduce algorithm consumption. In a more preferred embodiment of the present application, the spatial direction priority P of each spatial direction is calculated, wherein the spatial direction priority P and the time to collision (TTC) represent the predicted time of collision between the vehicle and the target in a certain direction. The heading angle cosine and the target category weight are positively correlated. For example, the lidar increases the pulse density in the direction with high spatial direction priority P value, and the camera increases the frame rate in this area. When a high-risk target such as a pedestrian crossing the road appears, the TTC is small and the category weight is high, the spatial direction priority P value will quickly increase, triggering real-time redistribution of sensor resources, ensuring that critical targets are not missed. The heading angle cosine cosθ represents the cosine value of the angle θ between the direction of the target and the vehicle heading angle, ranging from [-1, 1]. The target category weight W is a preset weight value according to the danger level of the target, based on prior rules or machine learning, for example: the W of the pedestrian = 1.0, which is the highest priority, and the collision consequence is the most serious, and the W of the non-motor vehicle = 0.8.

[0051] To ensure the overall system efficiency and energy consumption while significantly improving the point cloud quality in the key area, in a more preferred embodiment of the present application, in step S2, the pulse density ρ and / or laser energy E in the ROI direction is increased by a first preset proportion according to the spatial direction priority P value, wherein,

[0052] ∑ρ i ≤ρ total

[0053] ∑E i ≤E total

[0054] In the non-ROI direction, the pulse density ρ and / or laser energy E is reduced by a second preset proportion, so that the total pulse number ρ total and the total power consumption E total do not increase, wherein ρ i is the pulse density of the i-th direction, E iLaser energy for the i-th direction. If the ROI range is too large or the number is too large, it will still cause tight computing power; if the range is too small, it may miss the risk target and needs to be controlled through algorithms. For example, adaptive ROI scaling, priority sorting to achieve accurate control. Pulse density p is the number of laser pulses per unit angle, for example pulses / °, which determines the spatial resolution of the point cloud. The higher the p, the denser the point cloud. Laser energy E is the emission energy of a single laser, which affects the echo signal strength. The higher the E, the clearer the echo of the long-distance / low reflectivity target. For non-ROI directions, such as open road surfaces and target-free areas, reduce the pulse density p and / or laser energy E by a second preset ratio to avoid invalid energy consumption.

[0055] In order to realize the optimization of laser radar error from the hardware layer to the algorithm layer, while ensuring low cost, significantly improve the angle and distance accuracy of the point cloud, and provide more reliable underlying data support for the environment perception of autonomous driving, especially for high-precision perception demand in complex urban road conditions. When the vehicle is jolted, accelerated or decelerated, the IMU captures the attitude change in real time, and cooperates with the fixed reference of the corner reflector to quickly correct the scanning angle offset caused by mechanical vibration and improve the anti-interference performance. In a more preferred embodiment of the present application, an angle reflector and an inertial measurement unit are built into the shell of the laser radar, and a loop calibration is triggered once every preset time. The phase error and distance measurement error are calculated in real time using inertial measurement unit data and echo data; the phase error and distance measurement error are written into the FPGA lookup table for phase correction of the next synthesized frame. A high-reflectivity corner reflector is fixed on the inner wall of the laser radar shell to ensure that a fixed number of laser pulses are irradiated to the corner reflector during each scan. Preferably, 1 is set every 10°, for a total of 36, forming a self-echo reference point. The IMU is rigidly connected with the laser radar, and the real-time attitude of the laser radar is synchronously collected, and the sampling frequency is matched with the scanning frequency of the laser radar. Preferably, the sampling frequency is 1000 Hz. Loop calibration is triggered once every preset time, or triggered additionally when the laser radar is vibrating violently, to ensure the timeliness of calibration in dynamic scenes. For example, the preset time is 50 ms, i.e. 20 Hz. When the IMU detects that the acceleration exceeds the threshold, loop calibration is triggered. Look-up table (LUT) construction: a two-dimensional lookup table is established in the field-programmable gate array (FPGA), with the horizontal axis being the laser radar scanning angle 0°-360° and the vertical axis being the error value. After each loop calibration, the calculated error value is written into the LUT according to the angle index, overwriting the old value.

[0056] For example, in a high-speed straight-line scene

[0057] The intelligent vehicle travels at 100 km / h, and there is a 0.3 m obstacle 200 m ahead. Take N = 4,beam = 0.1° to θ eff = 0.025° so that the lateral resolution is about equal to 8.7cm, meeting the early warning requirement. The ROI direction pulse density is increased by 2 times, the laser energy is increased by 15%, and the average power consumption remains unchanged.

[0058] For example, in the left turn scene in the urban area

[0059] The speed of the oncoming electric vehicle is 30km / h, and the TTC is 1.8s; the ROI direction pulse density is increased by 2.5 times, the energy is increased by 20%, and the system triggers the brake 0.3s in advance.

[0060] Table 1 Comparison in different scenarios

[0061]

[0062] Obviously, the additional reaction time not only helps to reduce the possibility of accidents, but also greatly improves the comfort of driving.

[0063] The application also discloses a laser radar device based on the laser radar angle resolution enhancement method, and the laser radar device comprises,

[0064] A priority calculation unit is configured to calculate a spatial direction priority P of each spatial direction based on vehicle heading, road information and target information obtained by a vehicle external sensor;

[0065] A pulse energy distribution unit is configured to increase the pulse density ρ and / or the laser energy E of the ROI direction by a first preset proportion according to the value of the spatial direction priority P, and to decrease the pulse density ρ and / or the laser energy E of the non-ROI direction by a second preset proportion according to the value of the spatial direction priority P;

[0066] A phase offset multi-phase synthesis module is configured to synthesize the scanning data generated after N continuous scans into one synthesis frame, and to calculate the effective angle resolution;

[0067] A data interface module is configured to define and analyze an Ethernet frame; wherein the Ethernet frame format comprises: an Ethernet frame header, a self-defined header, and M groups of 4Byte data units, wherein M≤362;

[0068] An online self-calibration module is configured to perform loop calibration of the laser radar and phase correction of the next synthesis frame, and comprises an angle reflector and an inertial measurement unit, and the angle reflector comprises a spiral optical fiber loop.

[0069] The laser radar device comprises a priority calculation unit, which takes the target information obtained by the vehicle heading, road information and external sensors (such as a camera, a millimeter wave radar) as input, quantitatively calculates the spatial direction priority P of each spatial direction, and provides a decision basis for resource allocation. The pulse energy allocation unit dynamically adjusts the resources based on the priority P, increases the laser pulse density p and / or the laser energy E of the ROI with high priority by a first preset proportion, and reduces the corresponding resources by a second preset proportion in the non-ROI direction, so as to balance the energy efficiency and detection accuracy. The phase offset multi-phase synthesis module synthesizes the data of N continuous scanning into a single synthesis frame, improves the effective angular resolution, and breaks through the accuracy limit of single scanning. The data interface module defines and analyzes the Ethernet frame containing the Ethernet frame header, the custom header and M groups of 4Byte data units, and guarantees the efficient transmission and hardware compatibility of the point cloud data. The online self-calibration module completes the laser radar loop calibration by relying on the corner reflector containing the spiral optical fiber loop and the inertial measurement unit, calculates the phase and distance errors, and is used for phase correction of the next synthesis frame, and ensures the long-term measurement accuracy stability.

[0070] The present application provides an electronic device, at least one processor; and

[0071] A memory in communication connection with the at least one processor; wherein,

[0072] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above-mentioned laser radar angle resolution enhancement method.

[0073] The present application provides a machine readable storage medium, the machine readable storage medium stores instructions, the instructions are used to make the machine execute the laser radar angle resolution enhancement method of the present application as described above.

[0074] It should be noted that, for the above-mentioned various method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other order or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily necessary for the present application.

[0075] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0076] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0077] The above merely describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method of laser radar angle resolution enhancement, characterized by, The laser radar angle resolution enhancement method comprises: In step S1, the spatial direction priority P is calculated according to the vehicle heading, road information and target information obtained by the vehicle external sensor; wherein the vehicle external sensor at least includes a camera and a millimeter wave radar; In step S2, the laser pulse density and energy level are adjusted according to the spatial direction priority within the single circle scanning of the laser radar; Step S3, the scanning data generated after continuous scanning N circles as a composite frame, each circle starting angle offset θ beam / N, multi-phase filtering and coherent superposition in the digital domain to obtain effective angular resolution θ eff , wherein N is a positive integer greater than 1, θ beam is the inherent beam width of the laser radar.

2. The LIDAR angular resolution enhancement method of claim 1, wherein, In step S3, the point cloud data is obtained by performing multi-phase filtering and coherent superposition processing through FFT or MUSIC super-resolution algorithm.

3. The method of laser radar angle resolution enhancement according to claim 2, characterized in that, The point cloud data is twice interpolated to obtain effective angular resolution θ eff = θ beam / (2 N ).

4. The ladar angular resolution enhancement method of claim 1, wherein, The vehicle heading includes a heading angle; and the road information includes road curvature.

5. The ladar angular resolution enhancement method of claim 1, wherein, The spatial direction priority P of each spatial direction is calculated, wherein the spatial direction priority P is positively correlated with the time to collision TTC, the heading angle cosine and the target category weight.

6. The lidar angular resolution enhancement method of any of claims 1-5, wherein, In step S2, the pulse density ρ and / or laser energy E of the ROI direction are increased by a first preset proportion according to the spatial direction priority P value, wherein, ∑ρ i ≤ρ total ∑E i ≤E total decrease the pulse density p and / or the laser energy E in the non-ROI direction according to a second preset ratio, so that the total pulse number p total and the total power consumption E total do not increase, wherein p i is the pulse density of the i-th direction, and E i is the laser energy of the i-th direction.

7. The ladar angular resolution enhancement method of claim 6, wherein, The angle reflector and the inertial measurement unit are arranged in the shell of the laser radar, the phase error and the distance measurement error are calculated in real time by using the inertial measurement unit data and the echo data once every preset time, and the phase error and the distance measurement error are written into the FPGA lookup table for phase correction of the next synthesized frame.

8. A lidar device according to the lidar angle resolution enhancement method of any one of claims 1-7, characterized in that, The laser radar device comprises, A priority calculation unit for calculating the spatial direction priority P of each spatial direction according to the vehicle heading, road information and target information obtained by the vehicle external sensor; A pulse energy distribution unit for increasing the pulse density ρ and / or laser energy E of the ROI direction by a first preset proportion according to the spatial direction priority P value; and decreasing the pulse density ρ and / or laser energy E of the non-ROI direction by a second preset proportion according to the spatial direction priority P value; A phase offset multi-phase synthesis module for synthesizing the scanning data generated after N continuous scans into a synthesized frame and calculating the effective angular resolution; A data interface module for defining and analyzing the Ethernet frame; wherein the Ethernet frame format comprises: an Ethernet frame header, a self-defined header, M groups of 4 Byte data units, wherein M≤362; An online self-calibration module for the loop calibration of the laser radar and the phase correction of the next synthesized frame, comprising an angle reflector and an inertial measurement unit, and the angle reflector comprises a spiral optical fiber loop. At least one processor; 9. An electronic device, comprising: And A memory in communication connection with the at least one processor; wherein The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the laser radar angle resolution enhancement method in any one of claims 1-7. The machine readable storage medium stores instructions for causing a machine to execute the laser radar angle resolution enhancement method in any one of claims 1-7.

10. A machine-readable storage medium, characterized in that, ​

Images