Target recognition method, system and storage medium

Abstract

The present invention relates to the field of target recognition technology, and specifically provides a target recognition method, system, and storage medium, which are intended to solve the technical problem of inaccurate recognition results caused by inconsistent data acquisition times of multiple sensors in existing target recognition methods. To this end, the target recognition method of the present invention includes: synchronizing the time of all sensors based on a predetermined global clock; obtaining the data acquisition start time corresponding to each sensor based on preset acquisition synchronization parameters and acquisition mode information of each sensor; controlling each sensor to perform data acquisition on the target to be identified at the corresponding data acquisition start time, and obtaining multiple synchronous sampling data under the same time axis; and identifying the target to be identified based on the multiple synchronous sampling data. The technical solution provided by the present invention can obtain more accurate target recognition results.

Description

Technical Field

[0001] The present invention relates to the field of target recognition technology, and specifically provides a target recognition method, system and storage medium. Background Art

[0002] In current target recognition systems, when multiple sensors are deployed to collect data from the same target, data asynchrony often occurs. This is especially true for different sensor types. Because each sensor's scanning mechanism differs, the timing of sensor data collection is inconsistent. This results in the target being captured / scanned at different times by different sensors, and its corresponding position and size in the coordinate systems of each sensor are also inconsistent. Consequently, current target recognition is largely based on obtaining individual recognition results from the collected data of each sensor, and then fusing these results to produce the final target recognition result.

[0003] Practice has shown that the above scheme cannot accurately identify moving objects or small objects, resulting in many false positives and missed positives in actual applications. Based on this, a new target recognition scheme is needed in this field to solve the above problems. Summary of the Invention

[0004] In order to overcome the above-mentioned defects, the present invention is proposed to provide a solution or at least partially solve the technical problem of inaccurate recognition results caused by inconsistent data acquisition time of multiple sensors in existing target recognition methods.

[0005] In a first aspect, the present invention provides a target recognition method, which is applied to a target recognition system; the system comprises: at least two sensors for collecting data of a target to be recognized; the method comprises:

[0006] synchronizing time for all of the sensors based on a predetermined global clock;

[0007] Based on the preset acquisition synchronization parameters and the acquisition mode information of each sensor, obtaining the data acquisition start time corresponding to each sensor;

[0008] Controlling each of the sensors to collect data on the target to be identified at the corresponding data collection start time, and obtaining a plurality of synchronous sampling data on the same time axis;

[0009] The target to be identified is identified based on the multiple synchronous sampling data.

[0010] In one technical solution of the above target recognition method, the time synchronization of all the sensors based on a predetermined global clock includes:

[0011] Acquire the global clock as a master clock;

[0012] gPTP is used to synchronize the clock of each sensor with the master clock.

[0013] In one technical solution of the target recognition method, controlling each sensor to collect data on the target to be identified at the corresponding data collection start time to obtain multiple synchronous sampling data on the same time axis includes:

[0014] generating an acquisition trigger signal corresponding to each of the sensors based on the global clock and the acquisition mode information of each of the sensors;

[0015] The following operations are performed for each of the sensors: at the start time of data collection corresponding to the sensor, an acquisition trigger signal corresponding to the sensor is sent to the sensor, so that the sensor starts to collect data of the target to be identified based on the acquisition trigger signal, and obtains multiple synchronous sampling data under the same time axis.

[0016] In one technical solution of the target recognition method, the acquisition mode information includes an acquisition frequency; and generating an acquisition trigger signal corresponding to each sensor based on the global clock and the acquisition mode information of each sensor includes:

[0017] generating a synchronization signal based on the crystal oscillator frequency of the global clock;

[0018] The synchronization signal is frequency-divided by different multiples to generate an acquisition trigger signal corresponding to each of the sensors; wherein the frequency of the acquisition trigger signal corresponding to each of the sensors is consistent with the acquisition frequency of the sensor.

[0019] In one technical solution of the above-mentioned target recognition method, the target recognition system includes a fixed field terminal that interacts with the mobile terminal to be identified; the sensor includes: at least one of a camera, a lidar and a GNSS antenna installed at a predetermined position of the fixed field terminal; the preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment.

[0020] In one technical solution of the above target recognition method, the preset acquisition synchronization parameters further include: preset reference position coordinates; the method further includes:

[0021] Obtaining a starting scanning angle of each sensor based on the reference position coordinates and the acquisition mode information of each sensor;

[0022] And / or, controlling each of the sensors to collect data on the target to be identified at the corresponding data collection start time to obtain a plurality of synchronous sampling data on the same time axis, including:

[0023] Each sensor is controlled to collect data on the target to be identified at a corresponding starting scanning angle at a corresponding starting time of data collection, so as to obtain a plurality of synchronous sampling data on the same time axis.

[0024] In one technical solution of the above target recognition method, the reference position coordinate is at least one; the reference position coordinate is pre-set in the following manner:

[0025] Based on the installation position information of each sensor, obtaining the field of view angle information of each sensor;

[0026] Acquiring a predetermined ROI area in the fixed field terminal;

[0027] Based on the ROI area and the field of view angle information of each of the sensors, obtaining a coordinate range of at least one reference position and a sensor corresponding to each of the reference positions;

[0028] A reference position coordinate corresponding to each coordinate range is determined from the coordinate range; wherein each reference position coordinate satisfies the following condition: in a preset scanning cycle, the time stamp difference between the sensor corresponding to the reference position and the reference position is the smallest.

[0029] In one technical solution of the above target recognition method, obtaining the starting scanning angle of each sensor based on the reference position coordinates and the acquisition mode information of each sensor includes:

[0030] Converting the reference position coordinates into the coordinate system of each sensor to obtain reference coordinates corresponding to each sensor;

[0031] For each of the sensors, the following operations are performed: based on the reference coordinates corresponding to the sensor and the acquisition mode information of the sensor, a starting scanning angle of the sensor is obtained.

[0032] In one technical solution of the above target recognition method, the method further includes:

[0033] Testing the multiple synchronous sampling data under the same time axis;

[0034] When the test fails, the acquisition synchronization parameter is updated.

[0035] In one technical solution of the above target recognition method, the checking of the multiple synchronous sampling data under the same time axis includes:

[0036] For each synchronous sampling data, perform the following operations:

[0037] Calculating the time corresponding to the frame start position of the synchronous sampling data under the preset acquisition synchronization parameters based on the acquisition frequency of the sensor corresponding to the synchronous sampling data;

[0038] Determining whether the time corresponding to the frame start position of the synchronous sampling data matches the acquisition trigger signal corresponding to the sensor;

[0039] When there is a match, it is determined that the synchronous sampling data has passed the inspection.

[0040] In one technical solution of the above target recognition method, the preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment; and the checking of the multiple synchronous sampling data under the same time axis includes:

[0041] For each synchronous sampling data, perform the following operations:

[0042] Acquire the moment when the sensor corresponding to the synchronous sampling data scans to the synchronous scanning direction as the actual moment;

[0043] Calculating the difference between the actual time and the preset time;

[0044] Determining whether the difference is less than a preset deviation threshold;

[0045] When the difference is smaller than the preset deviation threshold, it is determined that the synchronous sampling data has passed the inspection.

[0046] In one technical solution of the above target recognition method, the target recognition system includes a fixed field terminal that exchanges information with the mobile terminal to be recognized; the method further includes:

[0047] Time synchronization is performed on the mobile terminal to be identified and the fixed field terminal.

[0048] In one technical solution of the above-mentioned target identification method, the time used by the fixed terminal includes: a preset first system time and a preset first local time; the time used by the mobile terminal to be identified includes: a preset second system time and a preset second local time; and the time synchronization of the mobile terminal to be identified and the fixed terminal includes:

[0049] Obtaining a time difference between the first system time and the second system time as a first time difference;

[0050] Obtaining a time difference between the first system time and the first local time as a second time difference;

[0051] Obtaining a time difference between the second system time and the second local time as a third time difference;

[0052] calculating a time difference between the first local time and the second local time based on the first time difference, the second time difference, and the third time difference;

[0053] Based on the time difference between the first local time and the second local time, time synchronization is performed on the mobile terminal to be identified and the fixed field terminal.

[0054] In one technical solution of the above target identification method, the time synchronization of the mobile terminal to be identified and the fixed field terminal further includes:

[0055] detecting a link delay between the fixed field terminal and the mobile terminal to be identified;

[0056] Determining whether the first time difference is credible based on the link delay and the first time difference;

[0057] The acquiring a time difference between the first system time and the first local time as a second time difference includes:

[0058] When the first time difference is credible, the time difference between the first system time and the first local time is obtained as the second time difference.

[0059] In a second aspect, a target recognition system is provided, which includes multiple sensors, a processor and a storage device; the sensors are used to collect data on the target to be identified; the storage device is suitable for storing multiple program codes, and the program codes are suitable for being loaded and run by the processor to execute the target recognition method described in any one of the technical solutions of the above-mentioned target recognition method.

[0060] In a third aspect, a computer-readable storage medium is provided, which stores a plurality of program codes, wherein the program codes are suitable for being loaded and run by a processor to execute the target recognition method described in any one of the technical solutions of the above-mentioned target recognition method.

[0061] Solution 1. A target recognition method, characterized in that it is applied to a target recognition system; the system comprises: at least two sensors for collecting data of the target to be recognized; the method comprises:

[0062] synchronizing time for all of the sensors based on a predetermined global clock;

[0063] Based on the preset acquisition synchronization parameters and the acquisition mode information of each sensor, obtaining the data acquisition start time corresponding to each sensor;

[0064] Controlling each of the sensors to collect data on the target to be identified at the corresponding data collection start time, and obtaining a plurality of synchronous sampling data on the same time axis;

[0065] The target to be identified is identified based on the multiple synchronous sampling data.

[0066] Solution 2. The target recognition method according to Solution 1, wherein the step of performing time synchronization on all the sensors based on a predetermined global clock comprises:

[0067] Acquire the global clock as a master clock;

[0068] gPTP is used to synchronize the clock of each sensor with the master clock.

[0069] Solution 3. The target recognition method according to Solution 1 is characterized in that controlling each sensor to collect data on the target to be identified at the corresponding data collection start time to obtain multiple synchronous sampling data on the same time axis includes:

[0070] generating an acquisition trigger signal corresponding to each of the sensors based on the global clock and the acquisition mode information of each of the sensors;

[0071] The following operations are performed for each of the sensors: at the start time of data collection corresponding to the sensor, an acquisition trigger signal corresponding to the sensor is sent to the sensor, so that the sensor starts to collect data of the target to be identified based on the acquisition trigger signal, so as to obtain multiple synchronous sampling data under the same time axis.

[0072] Solution 4. The target recognition method according to Solution 3, wherein the acquisition mode information includes an acquisition frequency; and generating an acquisition trigger signal corresponding to each sensor based on the global clock and the acquisition mode information of each sensor comprises:

[0073] generating a synchronization signal based on the crystal oscillator frequency of the global clock;

[0074] The synchronization signal is frequency-divided by different multiples to generate an acquisition trigger signal corresponding to each of the sensors; wherein the frequency of the acquisition trigger signal corresponding to each of the sensors is consistent with the acquisition frequency of the sensor.

[0075] Solution 5. The target recognition method according to any one of Solutions 1-4 is characterized in that the target recognition system includes a fixed field terminal that interacts with the mobile terminal to be identified; the sensor includes: at least one of a camera, a lidar and a GNSS antenna installed at a predetermined position of the fixed field terminal; the preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment.

[0076] Solution 6. The target recognition method according to Solution 5, wherein the preset acquisition synchronization parameters further include: preset reference position coordinates; and the method further includes:

[0077] Obtaining a starting scanning angle of each sensor based on the reference position coordinates and the acquisition mode information of each sensor;

[0078] And / or, controlling each of the sensors to collect data on the target to be identified at the corresponding data collection start time to obtain a plurality of synchronous sampling data on the same time axis, including:

[0079] Each sensor is controlled to collect data on the target to be identified at a corresponding starting scanning angle at a corresponding starting time of data collection, so as to obtain a plurality of synchronous sampling data on the same time axis.

[0080] Solution 7. The target recognition method according to Solution 6, characterized in that the reference position coordinate is at least one; the reference position coordinate is pre-set in the following manner:

[0081] Based on the installation position information of each sensor, obtaining the field of view angle information of each sensor;

[0082] Acquiring a predetermined ROI area in the fixed field terminal;

[0083] Based on the ROI area and the field of view angle information of each of the sensors, obtaining a coordinate range of at least one reference position and a sensor corresponding to each of the reference positions;

[0084] A reference position coordinate corresponding to each coordinate range is determined from the coordinate range; wherein each reference position coordinate satisfies the following condition: in a preset scanning cycle, the time stamp difference between the sensor corresponding to the reference position and the reference position is the smallest.

[0085] Solution 8. The target recognition method according to Solution 6, wherein obtaining a starting scanning angle of each sensor based on the reference position coordinates and the acquisition mode information of each sensor comprises:

[0086] Converting the reference position coordinates into the coordinate system of each sensor to obtain reference coordinates corresponding to each sensor;

[0087] For each of the sensors, the following operations are performed: based on the reference coordinates corresponding to the sensor and the acquisition mode information of the sensor, a starting scanning angle of the sensor is obtained.

[0088] Solution 9. The target recognition method according to Solution 4, further comprising:

[0089] Testing the multiple synchronous sampling data under the same time axis;

[0090] When the test fails, the acquisition synchronization parameter is updated.

[0091] Solution 10. The target recognition method according to Solution 9, wherein the checking of the multiple synchronous sampling data under the same time axis comprises:

[0092] For each synchronous sampling data, perform the following operations:

[0093] Calculating the time corresponding to the frame start position of the synchronous sampling data under the preset acquisition synchronization parameters based on the acquisition frequency of the sensor corresponding to the synchronous sampling data;

[0094] Determining whether the time corresponding to the frame start position of the synchronous sampling data matches the acquisition trigger signal corresponding to the sensor;

[0095] When there is a match, it is determined that the synchronous sampling data has passed the inspection.

[0096] Solution 11. The target recognition method according to Solution 9, wherein the preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment; and the checking of the multiple synchronous sampling data on the same time axis includes:

[0097] For each synchronous sampling data, perform the following operations:

[0098] Acquire the moment when the sensor corresponding to the synchronous sampling data scans to the synchronous scanning direction as the actual moment;

[0099] Calculating the difference between the actual time and the preset time;

[0100] Determining whether the difference is less than a preset deviation threshold;

[0101] When the difference is smaller than the preset deviation threshold, it is determined that the synchronous sampling data has passed the inspection.

[0102] Solution 12. The target recognition method according to Solution 1 is characterized in that the target recognition system includes: a fixed field terminal that exchanges information with the mobile terminal to be recognized; the method further includes:

[0103] Time synchronization is performed on the mobile terminal to be identified and the fixed field terminal.

[0104] Solution 13. The target identification method according to Solution 12, characterized in that the time used by the fixed terminal includes: a preset first system time and a preset first local time; the time used by the mobile terminal to be identified includes: a preset second system time and a preset second local time; and the time synchronization of the mobile terminal to be identified and the fixed terminal includes:

[0105] Obtaining a time difference between the first system time and the second system time as a first time difference;

[0106] Obtaining a time difference between the first system time and the first local time as a second time difference;

[0107] Obtaining a time difference between the second system time and the second local time as a third time difference;

[0108] calculating a time difference between the first local time and the second local time based on the first time difference, the second time difference, and the third time difference;

[0109] Based on the time difference between the first local time and the second local time, time synchronization is performed on the mobile terminal to be identified and the fixed field terminal.

[0110] Solution 14. The target identification method according to Solution 13, wherein the step of performing time synchronization on the mobile terminal to be identified and the fixed field terminal further comprises:

[0111] detecting a link delay between the fixed field terminal and the mobile terminal to be identified;

[0112] Determining whether the first time difference is credible based on the link delay and the first time difference;

[0113] The acquiring a time difference between the first system time and the first local time as a second time difference includes:

[0114] When the first time difference is credible, the time difference between the first system time and the first local time is obtained as the second time difference.

[0115] Solution 15. A target recognition system, characterized in that it includes multiple sensors, a processor and a storage device; the sensors are used to collect data on the target to be identified; the storage device is suitable for storing multiple program codes, and the program codes are suitable for being loaded and run by the processor to execute the target recognition method described in any one of Solutions 1 to 14.

[0116] Solution 16. A computer-readable storage medium storing a plurality of program codes, wherein the program codes are suitable for being loaded and run by a processor to execute the target recognition method according to any one of Solutions 1 to 14.

[0117] The above one or more technical solutions of the present invention have at least one or more of the following beneficial effects:

[0118] In the technical solution of the present invention, all sensors are time-synchronized based on a predetermined global clock, and based on preset acquisition synchronization parameters and the acquisition mode information of each sensor, the data acquisition start time corresponding to each sensor is obtained. Each sensor is controlled to acquire data of the target to be identified at its corresponding data acquisition start time, and multiple synchronous sampling data on the same time axis are obtained. Based on the multiple synchronous sampling data, the target to be identified is identified. When multiple sensors are provided, the multiple sensors can be controlled to perform synchronous data acquisition at a unified time, thereby obtaining multiple synchronous sampling data on the same time axis. The synchronous sampling data is used to perform target identification, which can solve the technical problem of inaccurate recognition results caused by inconsistent acquisition time of multiple sensors in the prior art. Compared with the prior art, the technical solution provided by the present invention can obtain more accurate target recognition results. BRIEF DESCRIPTION OF THE DRAWINGS

[0119] The disclosure of the present invention will be more easily understood with reference to the accompanying drawings. Those skilled in the art will readily appreciate that these drawings are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Furthermore, similar numbers in the drawings represent similar components, wherein:

[0120] Figure 1 This is a flow chart of the main steps of a target recognition method according to an embodiment of the present invention;

[0121] Figure 2 is a timing diagram of multiple sensor synchronization in one embodiment of the present invention;

[0122] Figure 3is a schematic diagram of reference position selection in one embodiment of the present invention;

[0123] Figure 4 is a timing diagram of multiple camera triggering in one embodiment of the present invention;

[0124] Figure 5 Schematic diagram of a laser radar data frame synchronization mechanism according to an embodiment of the present invention;

[0125] Figure 6 is a schematic diagram of data transmission between a battery swap station and an electric vehicle in one embodiment of the present invention;

[0126] Figure 7 It is a schematic diagram of the main structure block diagram of a target recognition device according to an embodiment of the present invention.

[0127] Reference Signs List :

[0128] 11: first time synchronization unit; 12: starting time acquisition unit;

[0129] 13: Control unit; 14: Identification unit. DETAILED DESCRIPTION

[0130] Some embodiments of the present invention are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0131] In the description of the present invention, "module" and "processor" may include hardware, software, or a combination of both. A module may include hardware circuitry, various suitable sensors, communication ports, and memory. It may also include software components, such as program code, or a combination of software and hardware. A processor may be a central processing unit, a microprocessor, an image processor, a digital signal processor, or any other suitable processor. A processor has data and / or signal processing capabilities. A processor may be implemented in software, hardware, or a combination of both. Non-transitory computer-readable storage media include any suitable medium capable of storing program code, such as magnetic disks, hard disks, optical disks, flash memory, read-only memory, random access memory, and the like. The term "A and / or B" refers to all possible combinations of A and B, such as only A, only B, or both A and B. The terms "at least one of A or B" or "at least one of A and B" have similar meanings to "A and / or B" and may include only A, only B, or both A and B. The singular forms "one" and "the" may also include the plural forms.

[0132] In current target recognition systems, when multiple sensors are set up to collect data from the same target to be identified, there is often a problem of asynchronous data collection. For example, battery swap stations used to replace batteries in electric vehicles are fixed infrastructure systems. As the demand for environmental safety in the battery swap process increases, the requirements for sensor field coverage and recognition of small objects are also increasing accordingly. In the past, battery swap stations were usually surrounded by cameras for security monitoring needs. Currently, there are plans to deploy multiple different types of sensors at battery swap stations to enhance the perception of the surrounding environment and vehicles.

[0133] In existing fixed-location battery swap stations and fixed-placement sensor deployment solutions, the different scanning mechanisms / scanning methods of each sensor result in inconsistent sensor data collection times. This results in moving objects being scanned at different times by different sensors, and their corresponding positions and sizes being inconsistent in the coordinate systems of different sensors. As a result, most perception and recognition solutions rely on target recognition based on the collected data of each sensor, and then a fusion verification of the various recognition results to obtain the final perception and recognition result. Because this solution cannot combine the characteristics of different types of sensors to more accurately identify and locate the same object, especially moving objects, the recognition accuracy for small objects such as small animals, cones, and triangular warning signs is low, resulting in many false alarms and missed alarms.

[0134] Taking the battery swap station as an example, the present invention designs a multi-sensor fusion solution with a simple configuration and high perception accuracy under fixed sensor positions. Without departing from the inventive concept of the present invention, the multi-sensor fusion solution of the present invention is not limited to battery swap stations, but can also be used in charging stations, storage stations, car washes, repair shops and other fixed places where multiple sensors can be installed and there are target recognition needs.

[0135] This solution uses a global clock to synchronize the perception and recognition system's timeline with the timelines of each sensor. It also configures the sensor's start scanning time based on each sensor's scanning mechanism / scanning method and monitors its operational stability to obtain synchronized sampling data of different types on the same timeline. Furthermore, considering the fixed position of the sensors, this solution incorporates sensor extrinsic parameters into the synchronization scheme and selects a common reference point based on each sensor's FOV (Field of View) coverage. This achieves temporal and spatial synchronization of multiple sensors, providing a foundation for subsequent perception and recognition system processing of multi-sensor data and enabling forward fusion of multi-sensor raw data for better object recognition. Due to the unstable and high average link latency of wireless links, this solution establishes a time synchronization mechanism between fixed-site terminals (such as battery swap stations) and mobile terminals (such as electric vehicles). This ensures that messages sent and received by both parties are converted based on the time synchronization results, converting the other party's timestamps to locally available timestamps. Compensating the data for time dimensions through software can eliminate instability caused by large data link latency.

[0136] See attached Figure 1 , Figure 1 The target recognition method of the present invention is applied to a target recognition system, which includes at least two sensors for collecting data of the target to be recognized. Figure 1 As shown, the target recognition method in the embodiment of the present invention mainly includes the following steps S101 to S104.

[0137] Step S101, performing time synchronization on all the sensors based on a predetermined global clock;

[0138] In this embodiment, the target recognition system includes: a fixed terminal for exchanging information with a mobile terminal to be identified; the mobile terminal to be identified may be a car or another mobile terminal to be identified. The fixed terminal may be a battery swap station for electric vehicles, a charging station for electric vehicles, or a storage station for electric vehicle batteries. The fixed terminal may also be a fixed location such as a car wash or repair shop capable of identifying and locating vehicles. The sensor includes: at least one of a camera, a lidar, and a GNSS antenna installed at a predetermined location on the fixed terminal.

[0139] The following takes the fixed terminal as a battery swap station and the mobile terminal to be identified as an electric vehicle as an example to illustrate the specific technical solution of this embodiment.

[0140] In one embodiment, the target to be identified may be an electric vehicle that is stationary or in motion. The electric vehicle includes an ADAS (Advanced Driving Assistance System) domain controller; the battery swap station includes a central control unit for performing various data processing. The global clock may be a clock generated by the above-mentioned central control unit based on a local crystal oscillator. In order to accurately and effectively achieve time synchronization of multiple sensors, the time synchronization of all the sensors based on a predetermined global clock described in this embodiment includes: obtaining the global clock as the master clock; and synchronizing the clock of each sensor with the master clock using gPTP.

[0141] In this embodiment, each sensor can be connected to the central control unit via a wireless network (such as Ethernet) or a hardwired connection. When connected via a wireless network, gPTP can be used to synchronize the clocks of each sensor with the master clock. When connected via a hardwired connection, a corresponding hardwired connection time synchronization scheme can be used for synchronization, directly generating the corresponding trigger signal based on the local global clock. This embodiment uses Ethernet connection as an example to describe the synchronization scheme between each sensor and the master clock.

[0142] In this embodiment, the central control unit of the battery swap station uses its own global clock generated based on the local crystal oscillator as the master clock, and uses the gPTP (generalized Precision Time Protocol) of the IEEE802.1AS protocol to synchronize the clocks of each camera and each lidar with the master clock. The central control unit uses the master clock generated based on the local crystal oscillator as the source of gPTP time synchronization, and issues synchronization signals and timestamps for synchronization to all sensors connected to it. The purpose of gPTP is to achieve time synchronization. After time synchronization is achieved, the sensors are all on a unified time axis, and there is a chance to achieve sensor acquisition synchronization. Therefore, the implementation of gPTP is the basis for sensor time synchronization on Ethernet. Only based on the gPTP time synchronization mechanism and associating the local timestamp with gPTP can the time axis of all sensors be unified, providing a basis for subsequent sensor synchronous trigger sampling.

[0143] Step S102, obtaining a data collection start time corresponding to each sensor based on preset collection synchronization parameters and collection mode information of each sensor;

[0144] In this embodiment, the preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment. The acquisition synchronization parameters are set based on user expectations.

[0145] For example, if the user expects all sensors to scan in the X direction at 8 o'clock, the preset time can be set to 8 o'clock, and the synchronous scanning direction can be set to the X direction.

[0146] In this embodiment, the sensor's data collection mode information includes the sensor's data collection frequency, rotation direction, rotation speed, etc. The central control unit of the battery swap station automatically calculates the data collection start time for each sensor based on the user-set data collection synchronization parameters and the data collection mode information for each sensor.

[0147] In one embodiment, the central control unit may also directly send the above-mentioned acquisition synchronization parameters to each sensor, and each sensor automatically adjusts its own data acquisition start time based on the acquisition synchronization parameters and its own acquisition mode.

[0148] Step S103, controlling each of the sensors to collect data on the target to be identified at the corresponding data collection start time, to obtain a plurality of synchronous sampling data on the same time axis;

[0149] In order to accurately and effectively control the data acquisition process of each sensor, the present embodiment controls each sensor to acquire data of the target to be identified at the corresponding data acquisition start time to obtain multiple synchronous sampling data on the same time axis, including: generating an acquisition trigger signal corresponding to each sensor based on the global clock and the acquisition mode information of each sensor; and performing the following operations for each sensor: sending the acquisition trigger signal corresponding to the sensor to the sensor at the data acquisition start time corresponding to the sensor, so that the sensor starts to acquire data of the target to be identified based on the acquisition trigger signal, and obtains multiple synchronous sampling data on the same time axis.

[0150] As described above, the acquisition mode information described in this embodiment includes an acquisition frequency. Generating an acquisition trigger signal corresponding to each sensor based on the global clock and the acquisition mode information of each sensor in this embodiment includes: generating a synchronization signal based on the crystal oscillator frequency of the global clock; and dividing the synchronization signal by different multiples to generate an acquisition trigger signal corresponding to each sensor. The frequency of the acquisition trigger signal corresponding to each sensor is consistent with the acquisition frequency of that sensor.

[0151] Specifically, the SOC (System On Chip) of the central control unit generates a high-frequency synchronization signal based on the local global clock, divides the synchronization signal by different multiples based on the frequency of the synchronization signal, and generates a trigger signal for triggering each sensor, that is, the above-mentioned acquisition trigger signal, and then sends the trigger signal to each sensor through a hard line or a wireless network (such as Ethernet) to enable each sensor to start data acquisition. It should be noted that due to the small transmission delay through hard line transmission, each sensor can start data acquisition immediately after receiving the trigger signal. When transmitting through a wireless network, the network delay problem needs to be considered. Each trigger signal will carry a corresponding timestamp. After receiving the trigger signal, each sensor will determine its specific data acquisition time based on the corresponding timestamp to avoid the influence of network delay on the synchronization effect.

[0152] In practical applications, since the position and scanning method of each sensor are fixed, the station-side sensor time synchronization solution can choose a relatively simple method to reduce system complexity. Based on the SOC's own time base, set the ptp timestamp to a whole 100 ms as the time synchronization point (or other arbitrary value), such as Figure 2 As shown by the second ptp time axis arrow in , it is the synchronization signal generated by the global clock.

[0153] Assume that the user expects the LiDAR and camera to scan in a preset direction at a preset time, for example, scanning in the X direction at 8 o'clock. Since each sensor has a fixed scanning method, the PTP timestamp corresponding to each sensor's scanning start time (i.e., the aforementioned data collection start time) can be reversed. For example, the LiDAR may start scanning at 010ms, and the camera may start scanning at 078ms. During the system initialization phase, the user-set acquisition synchronization parameters are passed through the configuration interface. Based on these acquisition synchronization parameters and each sensor's acquisition method, the system automatically calculates the time when each sensor starts data collection.

[0154] Furthermore, in one embodiment, in order to achieve a better synchronization effect for each sensor, the preset acquisition synchronization parameters also include: preset reference position coordinates. Then the target recognition method described in this embodiment also includes: obtaining the starting scanning angle of each sensor based on the reference position coordinates and the acquisition mode information of each sensor. Under this premise, the control of each sensor in this embodiment to collect data on the target to be identified at the corresponding data acquisition start time and obtain multiple synchronous sampling data on the same time axis includes: controlling each sensor to collect data on the target to be identified at the corresponding data acquisition start time and with the corresponding starting scanning angle, and obtaining multiple synchronous sampling data on the same time axis.

[0155] That is, in this embodiment, the starting scanning angle of each sensor is obtained to achieve a better synchronization effect.

[0156] Furthermore, in one embodiment, in order to accurately obtain the starting scanning angle of each sensor, the starting scanning angle of each sensor is obtained based on the reference position coordinates and the acquisition method information of each sensor, including: converting the reference position coordinates into the coordinate system of each sensor to obtain the reference coordinates corresponding to each sensor; for each sensor, performing the following operations: obtaining the starting scanning angle of the sensor based on the reference coordinates corresponding to the sensor and the acquisition method information of the sensor.

[0157] In this embodiment, the reference position coordinates are rotated and translated based on the external and internal parameters of each sensor, converting them to the coordinate system of each sensor to obtain the reference coordinates corresponding to each sensor. For each sensor, the starting scanning angle can be calculated based on the reference coordinates and the sensor's data acquisition method information.

[0158] Furthermore, in one embodiment, in order to enable each sensor to further achieve a better synchronization effect, at least one of the reference position coordinates is pre-set in the following manner: based on the installation position information of each sensor, the field of view angle information of each sensor is obtained; a predetermined ROI area in the battery swap station is obtained; based on the ROI area and the field of view angle information of each sensor, a coordinate range of at least one reference position and a sensor corresponding to each reference position are obtained; from each coordinate range, a reference position coordinate corresponding to the coordinate range is determined; wherein each reference position coordinate satisfies the following conditions: in a preset scanning cycle, the timestamp difference between the sensor corresponding to the reference position and the reference position is the smallest.

[0159] The minimum timestamp difference means that each sensor can scan the reference position almost simultaneously. As each sensor continues scanning, it gradually moves away from the reference position and spreads toward the edge of the scanned image. The time difference between the images scanned by each sensor also gradually increases. This is consistent with the expectation in real applications that the error at the focus position (i.e., the reference position) is small, while the error at the edge of the image is large.

[0160] The ROI (Region of Interest) area is the area of ​​interest of the central control unit of the battery swap station when identifying the target object. Based on the ROI area and the field of view of each sensor, the overlapping area containing the ROI area in the field of view of each sensor can be obtained, and then the coordinate ranges of multiple reference positions can be obtained based on the overlapping area. Based on the conditions that the above-mentioned reference position coordinates should meet, a reference position coordinate can be determined from each coordinate range. For each reference position, a sensor within a preset range centered on the reference position can be selected as the sensor corresponding to the reference position.

[0161] Specifically, considering that each sensor has a different field of view and coverage, in order to achieve the best synchronization effect, multiple reference positions (i.e. Figure 3 Taking the battery swap station as an example, this embodiment selects multiple reference points based on the station-side coordinate system, such as Figure 3 shown. Figure 3 Ref M is the reference point inside the station, and Ref P is the reference point outside the station. Based on the sensor position calibration parameters, the reference point coordinates, and their own scan point cloud (i.e., the point cloud formed by the data collected by the lidar), the two lidars inside the station can calculate the data frame start time for subsequent synchronization verification. Assuming the sensor scanning method is fixed, the preset time, the synchronous scanning direction of all sensors corresponding to the preset time, and the starting scanning angle of each sensor are uniformly input into the sensor configuration to ensure that the synchronization effect of each sensor under the default scanning method meets the recognition accuracy requirements.

[0162] It should be noted that Figure 3 This is only a schematic diagram of the reference points, and the actual positions and numbers of the reference points are not specifically limited in this embodiment.

[0163] Furthermore, in one embodiment, in order to ensure the synchronization effect of each sensor, the target recognition method further includes: testing the multiple synchronous sampling data under the same time axis; when the test fails, updating the acquisition synchronization parameter.

[0164] Furthermore, in one embodiment, to accurately and effectively verify the synchronization effect, the verification of the multiple synchronous sampling data on the same time axis includes the following operations: for each synchronous sampling data, calculating the time corresponding to the frame start position of the synchronous sampling data under the preset acquisition synchronization parameters based on the acquisition frequency of the sensor corresponding to the synchronous sampling data; determining whether the time corresponding to the frame start position of the synchronous sampling data matches the acquisition trigger signal corresponding to the sensor; if so, determining that the synchronous sampling data has passed the verification. Correspondingly, if not, determining that the synchronous sampling data has failed the verification.

[0165] As described above, the preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment. Under this premise, in order to accurately and effectively verify the synchronization effect, the verification of multiple synchronous sampling data on the same time axis described in this embodiment includes: for each synchronous sampling data, the following operations are performed: obtaining the moment when the sensor corresponding to the synchronous sampling data scans to the synchronous scanning direction as the actual moment; calculating the difference between the actual moment and the preset moment; determining whether the difference is less than a preset deviation threshold; and determining that the synchronous sampling data has passed the verification when the difference is less than the preset deviation threshold.

[0166] That is, in this embodiment, the above two testing methods can be used to effectively test the synchronization effect of the sensor.

[0167] Specifically, the frame start position of the synchronous sampling data collected by the sensor under the current preset acquisition synchronization parameters is calculated based on the sensor's acquisition frequency, and the time corresponding to the frame start position is obtained to determine whether the time matches the acquisition trigger signal corresponding to the sensor. When it matches, it is determined that the synchronous sampling data test has passed, that is, it is determined that each sensor is synchronized. When each sensor is synchronized, the frame start time of the data collected by each sensor is stable, and the reference position can be automatically synchronized. The synchronization test method can also be to analyze the data collected by the sensor at regular intervals (1 minute or 5 minutes, depending on the stability of the sensor), obtain the time when the sensor scans to the above-mentioned preset synchronous scanning direction, and calculate the difference between this time and the preset time expected by the user, and determine whether the difference is within the allowable numerical range.

[0168] This embodiment unifies the time of each sensor based on a global clock, avoiding mutual coupling between multiple sensors. In practical applications, a dedicated time synchronization module can be set up to regularly verify the collected data based on the timestamps of the feedback data from each sensor and determine whether the collection synchronization parameters need to be updated to achieve time alignment of the data collected by each sensor.

[0169] In practical applications, for different camera types, the chip's internal hardware counter can be used to periodically trigger the synchronization signal based on the acquisition synchronization parameters. The video stream receiver records the timestamp based on the data frame start position and the timestamp that triggered the synchronization signal. The time synchronization module periodically verifies the timestamp output by the camera. If the camera synchronization error exceeds the set value, it outputs a phase adjustment value for the synchronization signal based on the deviation value and updates the hardware counter Count value. This ensures that the camera is synchronized with the global clock after the adjustment, achieving decoupling from other sensors.

[0170] Figure 4 FIG. 4 is a timing diagram of triggering multiple cameras in this embodiment, showing the timing of triggering the cameras to work. Figure 4 The two parallelograms in the figure represent the timing of camera exposure. Each time a trigger signal is received, the camera will delay for a period of time before performing the exposure operation. After the exposure operation, the corresponding data will be read out. The reference exposure center point refers to the visual center of the camera, which is located at the center of each parallelogram. At this center position, assuming that the exposure start time is 500ms, this solution can align the exposure start time of each camera to 500ms. For a camera with a frequency of 10Hz, the exposure time of the next frame should be 600ms. In this way, the time alignment operation of multiple cameras is achieved, thereby realizing the synchronous acquisition of multiple cameras.

[0171] In practical applications, for LiDAR, the preset acquisition synchronization parameters are all transmitted through Ethernet, and the LiDAR is required to achieve time synchronization with the SOC by supporting the gPTP protocol. Due to the uncertainty of data link delay during Ethernet transmission, thanks to the small change in the relative position of its own posture and reference point in the fixed-end scenario, it does not require high real-time dynamic adjustment of the above acquisition synchronization parameters. It only requires the LiDAR scanning mode to be fixed and the motor rotation direction and speed to be stable, and regular monitoring on the control end to achieve a better synchronization effect. However, due to the certain regularity of the distribution of data under the LiDAR scanning mode, the timestamps of data in different areas around the reference point scanned by multiple LiDARs in the fixed-end scenario are certain. Therefore, the design of this solution needs to consider the differences in the timestamps of the ROI areas in different rotation directions. Figure 5For example, through two reference points Ref M and Ref P, the area inside and outside the battery swap station is divided into four blocks. Different rotation directions are set for the lidar. The difference distribution of the two lidar data times in each area is different. It is necessary to ensure that the timestamp difference of the perceived ROI area is minimized, that is, to ensure that in the same scanning cycle, the timestamp of the reference position (i.e. Figure 5 The time stamp difference between the sensor corresponding to the reference point in the image and the reference position is the smallest, thus achieving the best data fusion effect.

[0172] exist Figure 5 In the figure, Ref M1 is for comparison with Ref M. As can be seen from the figure, the time difference between radars C and D scanning Ref M is 0, that is, radars C and D can scan Ref M simultaneously. However, there is a certain time difference between radars C and D scanning Ref M1, that is, radars C and D cannot scan Ref M1 simultaneously. Therefore, Ref M is the optimal reference position corresponding to radars C and D. Similarly, Ref P1 is for comparison with Ref P. As can be seen from the figure, the time difference between radars A and B scanning Ref P is 0, that is, radars A and B can scan Ref P simultaneously. However, there is a certain time difference between radars A and B scanning Ref P1, that is, radars A and B cannot scan Ref P1 simultaneously. Therefore, Ref P is the optimal reference position corresponding to radars A and B.

[0173] Step S104: identifying the target to be identified based on the multiple synchronous sampling data.

[0174] The above steps S101-S104, the present invention can solve the technical problem of inaccurate recognition results caused by inconsistent data collection time of multiple sensors in existing target recognition methods. Furthermore, for the mobile terminal to be identified and the fixed field terminal, the mobile terminal to be identified and the fixed field terminal need time synchronization to ensure that the data transmission at both ends will not be degraded in real time due to link delay. If there is no time synchronization, the data link delay of 10ms->100ms->1s has a direct impact on the business. If the two ends achieve time synchronization, the data transmitted between the two parties is accompanied by a local timestamp, and can be converted into the timestamp of the other end through the time synchronization error. Although there is a certain delay after the other end receives the data, it can still match the data under the local corresponding timestamp according to the timestamp, thereby achieving better recognition and control accuracy, greatly reducing the impact of data link delay uncertainty on the business, and greatly reducing the dependence of data services on network delays.

[0175] Therefore, in order to achieve the above-mentioned effect and obtain more accurate target recognition results, the target recognition method described in one embodiment of the present invention also includes: time synchronization of the mobile terminal to be identified and the fixed field terminal. In this embodiment, the time used by the fixed field terminal includes: a preset first system time and a preset first local time, that is, different business modules in the fixed field terminal use different time bases. Depending on the business, some modules use the first system time, and some modules use the first local time. Similarly, the time used by the mobile terminal to be identified includes: a preset second system time and a preset second local time, that is, different business modules in the mobile terminal to be identified also use different time bases. Some modules use the second system time, and some modules use the second local time. Under this premise, the time synchronization of the mobile terminal to be identified and the fixed field terminal described in this embodiment includes: obtaining the time difference between the first system time and the second system time as the first time difference; obtaining the time difference between the first system time and the first local time as the second time difference; obtaining the time difference between the second system time and the second local time as the third time difference; calculating the time difference between the first local time and the second local time based on the first time difference, the second time difference and the third time difference; and performing time synchronization on the mobile terminal to be identified and the fixed field terminal based on the time difference between the first local time and the second local time.

[0176] The following still takes the fixed terminal as a battery swap station and the mobile terminal to be identified as an electric vehicle as an example to illustrate the technical solution provided by this embodiment.

[0177] Specifically, the above synchronization method can be used to synchronize the time between the electric vehicle and the battery swap station. In one embodiment, based on the optimized Linux clockdiff command, the time difference between the first system time and the second system time is detected as the first time difference; the time difference between the first system time and the first local time is obtained as the second time difference; the time difference between the second system time and the second local time is obtained as the third time difference; based on the first time difference, the second time difference, and the third time difference, the time difference between the first local time and the second local time is calculated; and based on the time difference between the first local time and the second local time, the time synchronization between the electric vehicle and the battery swap station is performed.

[0178] The Linux clockdiff command used in this embodiment is an optimized Linux clockdiff command, which has been optimized in terms of the number of tests, test time, and calculation logic of the system deviation, so as to obtain a more accurate time deviation between the electric vehicle and the battery swap station, thereby achieving a better time synchronization effect between the two.

[0179] Furthermore, in one embodiment, in order to achieve a more accurate synchronization effect, the time synchronization of the electric vehicle and the battery swap station also includes: detecting the link delay between the battery swap station and the electric vehicle; judging whether the first time difference is credible based on the link delay and the first time difference; under this premise, the acquisition of the time difference between the first system time and the first local time as the second time difference described in this embodiment includes: when the first time difference is credible, acquiring the time difference between the first system time and the first local time as the second time difference.

[0180] Specifically, the time synchronization solution between the battery swap station and the electric vehicle in this embodiment relies on the clockdiff tool and the ping command in the Linux tool. For the time synchronization between the mobile end and the fixed end of the data transmission, a transponder module is arranged at each end, such as Figure 6 As shown, two repeater modules perform timestamp conversion on the information transmitted at both ends. After the connection is established between the two parties, a clockdiff test is initiated through the repeater module, and the link delay of both parties is continuously tested using the ping command. Combined with the system time deviation at both ends obtained by clockdiff and the ping command result, it is determined whether the current clockdiff result is credible. If it is credible, the difference between the local time and system time of the fixed end, and the difference between the local time and system time of the mobile end are calculated and sent to the fixed end or the mobile end to further calculate the time difference between the local time of both ends, thereby synchronizing the time of both ends. The repeaters at both ends are arranged symmetrically and have the same logic. Both parties can obtain the timestamp error at both ends in this way. In addition, the symmetrical arrangement can ensure the reuse of the control logic at both ends, and it can also be considered to perform timestamp conversion for uplink data and downlink data at one end. The repeater can judge the link status to determine the correctness of the time synchronization result, and send the above link status and the determined result synchronously to the data consumer, that is Figure 6 Each application module in.

[0181] In addition to time synchronization with electric vehicles, fixed field terminals (such as battery swap stations) can also synchronize time with other mobile devices, such as mobile phones and tablets, to ensure the consistency of timestamps at both ends during data transmission.

[0182] The technical solution provided by the embodiment of the present invention synchronizes all sensors based on a predetermined global clock, obtains the data collection start time corresponding to each sensor based on preset collection synchronization parameters and the collection mode information of each sensor, controls each sensor to collect data of the target to be identified at its corresponding data collection start time, obtains multiple synchronous sampling data on the same time axis, and identifies the target to be identified based on the multiple synchronous sampling data. When multiple sensors are provided, the multiple sensors can be controlled to perform synchronous data collection at a unified time, thereby obtaining multiple synchronous sampling data on the same time axis, and using the synchronous sampling data for target identification. This can solve the technical problem of inaccurate recognition results caused by inconsistent data collection times of multiple sensors in the prior art. Compared with the prior art, the technical solution provided by the present invention can obtain more accurate target identification results.

[0183] This invention utilizes a global clock within a computing unit at a fixed station (battery swap station) to actively trigger multiple sensors, such as cameras and lidar. Based on the different scanning modes and operating principles of each sensor, corrections are applied at different times based on a unified time reference, enabling synchronous data acquisition and spatial alignment across multiple sensors. This solves existing issues such as inconsistent timelines across multiple sensors, inconsistent data sampling times due to different scanning modes, and the inability to align data images within a common reference frame due to differing coordinate systems across sensors.

[0184] This method deploys transponder modules at both the fixed station (battery swap station) and the mobile terminal (electric vehicle), utilizing an improved version of the Linux clockdiff tool to obtain the time difference between the two systems and perform conversion based on the difference in the local time axis. This solves the problem of large recognition errors and poor control accuracy caused by long and unstable data link delays when wirelessly connecting fixed stations and mobile terminals.

[0185] The present invention provides a simple and robust technical solution for multi-sensor synchronization based on the arrangement of sensors at fixed positions in the battery swap station scenario, providing a better technical foundation for sensor perception fusion. It solves the problem that in the multi-sensor fusion process, the image appears motion blurred / the key features of the image are lost and elongated / the recognition results of multiple sensors are inconsistent due to the asynchronous sampling time and the asynchronous spatial coordinates, which leads to the misidentification or missed recognition of the target by the perception system. In addition, a time synchronization method is provided for the communication interconnection solution between the computing platform at the battery swap station end and electric vehicles and other mobile devices such as mobile phones. The clockdiff tool under Linux is used to help the mobile end and the station end calculate the time error between the two systems, ensure the consistency of the timestamps of the two systems during the data transmission process, reduce the dependence on data link delay in autonomous driving, avoid business interruption caused by unstable wireless communication links between the mobile end and the station end and large delay jitter caused by environmental interference, and provide a basis for communication between the mobile end and the station end.

[0186] It should be pointed out that although the various steps in the above embodiments are described in a specific order, those skilled in the art will understand that in order to achieve the effects of the present invention, different steps do not have to be performed in such an order. They can be performed simultaneously (in parallel) or in other orders. These changes are within the scope of protection of the present invention.

[0187] Furthermore, the present invention also provides a target recognition device.

[0188] See attached Figure 7 , Figure 7 This is a block diagram of the main structure of a target recognition device according to an embodiment of the present invention. The device is applied to a target recognition system, which includes at least two sensors for collecting data of the target to be recognized. Figure 7 As shown, the target recognition device in the embodiment of the present invention mainly includes a first time synchronization unit 11, a start time acquisition unit 12, a control unit 13 and a recognition unit 14.

[0189] A first time synchronization unit 11 is configured to synchronize the time of all the sensors based on a predetermined global clock;

[0190] A start time acquisition unit 12 is configured to obtain a data collection start time corresponding to each sensor based on preset collection synchronization parameters and collection mode information of each sensor;

[0191] The control unit 13 is configured to control each of the sensors to collect data on the target to be identified at the corresponding data collection start time, thereby obtaining a plurality of synchronous sampling data on the same time axis;

[0192] The identification unit 14 is configured to identify the target to be identified based on the plurality of synchronous sampling data.

[0193] In this embodiment, the first time synchronization unit 11 performs time synchronization on all sensors in the following manner:

[0194] Acquire the global clock as a master clock;

[0195] gPTP is used to synchronize the clock of each sensor with the master clock.

[0196] In this embodiment, the control unit 13 includes:

[0197] a trigger signal generating unit, configured to generate an acquisition trigger signal corresponding to each of the sensors based on the global clock and the acquisition mode information of each of the sensors;

[0198] The trigger signal sending unit is used to perform the following operations for each of the sensors: at the start time of data collection corresponding to the sensor, the collection trigger signal corresponding to the sensor is sent to the sensor, so that the sensor starts to collect data of the target to be identified based on the collection trigger signal, and obtains multiple synchronous sampling data under the same time axis.

[0199] In this embodiment, the acquisition mode information includes: acquisition frequency; the trigger signal generating unit generates an acquisition trigger signal corresponding to each sensor in the following manner:

[0200] generating a synchronization signal based on the crystal oscillator frequency of the global clock;

[0201] The synchronization signal is frequency-divided by different multiples to generate an acquisition trigger signal corresponding to each of the sensors; wherein the frequency of the acquisition trigger signal corresponding to each of the sensors is consistent with the acquisition frequency of the sensor.

[0202] In this embodiment, the target recognition system includes: a battery swap station for replacing batteries of electric vehicles; the sensors include: a camera and a lidar installed at a predetermined position of the battery swap station; the preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment.

[0203] In this embodiment, the preset acquisition synchronization parameters further include: preset reference position coordinates. The target recognition device of this embodiment further includes:

[0204] a starting scanning angle acquiring unit, configured to acquire a starting scanning angle of each of the sensors based on the reference position coordinates and the acquisition mode information of each of the sensors;

[0205] The control unit 13 is further configured to control each sensor to collect data on the target to be identified at a corresponding starting scanning angle at a corresponding starting moment of data collection, so as to obtain a plurality of synchronous sampling data on the same time axis.

[0206] In this embodiment, the reference position coordinate is at least one; the reference position coordinate is pre-set in the following manner:

[0207] Based on the installation position information of each sensor, obtaining the field of view angle information of each sensor;

[0208] Obtaining a predetermined ROI area in the battery swap station;

[0209] Based on the ROI area and the field of view angle information of each of the sensors, obtaining a coordinate range of at least one reference position and a sensor corresponding to each of the reference positions;

[0210] A reference position coordinate corresponding to each coordinate range is determined from the coordinate range; wherein each reference position coordinate satisfies the following condition: in a preset scanning cycle, the time stamp difference between the sensor corresponding to the reference position and the reference position is the smallest.

[0211] In this embodiment, the starting scanning angle obtaining unit obtains the starting scanning angle of each sensor in the following manner:

[0212] Converting the reference position coordinates into the coordinate system of each sensor to obtain reference coordinates corresponding to each sensor;

[0213] For each of the sensors, the following operations are performed: based on the reference coordinates corresponding to the sensor and the acquisition mode information of the sensor, a starting scanning angle of the sensor is obtained.

[0214] Furthermore, the device described in this embodiment also includes:

[0215] A testing unit, configured to test the plurality of synchronous sampling data under the same time axis;

[0216] An updating unit is configured to update the acquisition synchronization parameter when the test fails.

[0217] In this embodiment, the verification unit verifies the multiple synchronous sampling data under the same time axis in the following manner:

[0218] For each synchronous sampling data, perform the following operations:

[0219] Calculating the time corresponding to the frame start position of the synchronous sampling data under the preset acquisition synchronization parameters based on the acquisition frequency of the sensor corresponding to the synchronous sampling data;

[0220] Determining whether the time corresponding to the frame start position of the synchronous sampling data matches the acquisition trigger signal corresponding to the sensor;

[0221] When there is a match, it is determined that the synchronous sampling data has passed the inspection.

[0222] In this embodiment, the preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment; the verification unit verifies the multiple synchronous sampling data under the same time axis in the following manner:

[0223] For each synchronous sampling data, perform the following operations:

[0224] Acquire the moment when the sensor corresponding to the synchronous sampling data scans to the synchronous scanning direction as the actual moment;

[0225] Calculating the difference between the actual time and the preset time;

[0226] Determining whether the difference is less than a preset deviation threshold;

[0227] When the difference is smaller than the preset deviation threshold, it is determined that the synchronous sampling data has passed the inspection.

[0228] In this embodiment, the target recognition system includes: a battery swap station for replacing batteries in electric vehicles; the device described in this embodiment also includes:

[0229] The second time synchronization unit is used to synchronize the time between the electric vehicle and the battery swap station.

[0230] In this embodiment, the time used by the battery swap station includes: a first system time and a first local time; the time used by the electric vehicle includes: a second system time and a second local time; the second time synchronization unit includes:

[0231] a first time difference acquiring unit, configured to detect, based on a Linux clockdiff command, a time difference between the first system time and the second system time as a first time difference;

[0232] a second time difference acquiring unit, configured to acquire a time difference between the first system time and the first local time as a second time difference;

[0233] a third time difference acquiring unit, configured to acquire a time difference between the second system time and the second local time as a third time difference;

[0234] a calculation unit, configured to calculate a time difference between the first local time and the second local time based on the first time difference, the second time difference, and the third time difference;

[0235] A time synchronization subunit is used to synchronize the time of the electric vehicle and the battery swap station based on the time difference between the first local time and the second local time.

[0236] Furthermore, in this embodiment, the second time synchronization unit further includes:

[0237] a link delay detection unit, configured to detect a link delay between the battery swap station and the electric vehicle;

[0238] a judging unit, configured to judge whether the first time difference is credible based on the link delay and the first time difference;

[0239] The second time difference acquiring unit is further configured to acquire, when the first time difference is credible, a time difference between the first system time and the first local time as a second time difference.

[0240] In some embodiments, one or more of the first time synchronization unit 11, the start time acquisition unit 12, the control unit 13, and the identification unit 14 can be combined into one module. In one embodiment, the specific implementation functions can be described in steps S101-S104.

[0241] The target recognition device is used to perform Figure 1 The target recognition method embodiments shown in the figure are similar in their technical principles, technical problems solved and technical effects produced. Technicians in this technical field can clearly understand that for the convenience and conciseness of description, the specific working process and related instructions of the target recognition device can refer to the contents described in the embodiment of the target recognition method, and will not be repeated here.

[0242] In an embodiment of the present invention, the device may be a control device formed by various electronic devices. In some possible implementations, the device may include multiple storage devices and multiple processors. The program for executing the target recognition method of the above method embodiment can be divided into multiple subroutines, and each subroutine can be loaded and run by a processor to execute different steps of the target recognition method of the above method embodiment. Specifically, each subroutine can be stored in different memories respectively, and each processor can be configured to execute the program in one or more memories to jointly implement the target recognition method of the above method embodiment, that is, each processor executes different steps of the target recognition method of the above method embodiment respectively to jointly implement the target recognition method of the above method embodiment.

[0243] The aforementioned multiple processors may be processors deployed on the same device. For example, the aforementioned computer device may be a high-performance device composed of multiple processors, and the aforementioned multiple processors may be processors configured on the high-performance device. Furthermore, the aforementioned multiple processors may also be processors deployed on different devices. For example, the aforementioned computer device may be a server cluster, and the aforementioned multiple processors may be processors on different servers in the server cluster.

[0244] Those skilled in the art will appreciate that all or part of the processes in the method for implementing the above-mentioned embodiment of the present invention may also be accomplished by instructing the relevant hardware through a computer program. The computer program may be stored in a computer-readable storage medium. When the computer program is executed by a processor, it may implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which may be in source code form, object code form, executable file, or some intermediate form. The computer-readable storage medium may include: any entity or device, medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory, random access memory, electric carrier signal, telecommunication signal, and software distribution medium capable of carrying the computer program code. It should be noted that the content contained in the computer-readable storage medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electric carrier signals and telecommunication signals.

[0245] Furthermore, the present invention also provides a target recognition system. In an embodiment of a target recognition system according to the present invention, the target recognition system includes a plurality of sensors, a processor, and a storage device; the sensors are used to collect data of the target to be recognized; the storage device can be configured to store a program for executing the target recognition method of the above method embodiment, and the processor can be configured to execute the program in the storage device, which includes but is not limited to a program for executing the target recognition method of the above method embodiment. For ease of explanation, only the parts related to the embodiment of the present invention are shown. For specific technical details not disclosed, please refer to the method part of the embodiment of the present invention. The control device can be a control device device formed by various electronic devices.

[0246] Furthermore, the present invention also provides a computer-readable storage medium. In one embodiment of a computer-readable storage medium according to the present invention, the computer-readable storage medium can be configured to store a program for executing the target recognition method of the above-mentioned method embodiment, and the program can be loaded and run by a processor to implement the above-mentioned target recognition method. For ease of explanation, only the parts related to the embodiment of the present invention are shown. For specific technical details not disclosed, please refer to the method part of the embodiment of the present invention. The computer-readable storage medium can be a storage device formed by various electronic devices. Optionally, the computer-readable storage medium in the embodiment of the present invention is a non-transitory computer-readable storage medium.

[0247] Furthermore, it should be understood that since the configuration of each module is merely for the purpose of illustrating the functional units of the apparatus of the present invention, the physical devices corresponding to these modules may be the processor itself, or a portion of the software in the processor, a portion of the hardware, or a combination of software and hardware. Therefore, the number of modules in the figure is merely illustrative.

[0248] Those skilled in the art will appreciate that the various modules in the device can be adaptively split or merged. Such splitting or merging of specific modules does not cause the technical solution to deviate from the principles of the present invention. Therefore, the technical solutions after splitting or merging will fall within the scope of protection of the present invention.

[0249] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. A target recognition method, characterized in that: Applied to target recognition systems; The system includes: at least two sensors for collecting data on a target to be identified; the method includes: synchronizing time for all of the sensors based on a predetermined global clock; Obtaining a data acquisition start time corresponding to each sensor based on preset acquisition synchronization parameters and acquisition mode information of each sensor, wherein the preset acquisition synchronization parameters include: at least one preset time, a synchronous scanning direction of all the sensors corresponding to each preset time, and preset reference position coordinates; Obtaining a starting scanning angle of each sensor based on the reference position coordinates and the acquisition mode information of each sensor; Controlling each of the sensors to collect data on the target to be identified at a corresponding starting scanning angle at a corresponding starting time of data collection, to obtain a plurality of synchronous sampling data on the same time axis; Identifying the target to be identified based on the multiple synchronous sampling data; The reference position coordinates are preset in the following manner: Based on the installation position information of each sensor, obtaining the field of view angle information of each sensor; Obtaining a predetermined ROI area in a fixed field terminal; Based on the ROI area and the field of view angle information of each of the sensors, obtaining a coordinate range of at least one reference position and a sensor corresponding to each of the reference positions; A reference position coordinate corresponding to each coordinate range is determined from the coordinate range; wherein each reference position coordinate satisfies the following condition: in a preset scanning cycle, the time stamp difference between the sensor corresponding to the reference position and the reference position is the smallest.

2. The target recognition method according to claim 1, characterized in that: The step of performing time synchronization on all the sensors based on a predetermined global clock includes: Acquire the global clock as a master clock; gPTP is used to synchronize the clock of each sensor with the master clock.

3. The target recognition method according to claim 1, characterized in that: The controlling each sensor to collect data on the target to be identified at the corresponding data collection start time to obtain a plurality of synchronous sampling data on the same time axis includes: generating an acquisition trigger signal corresponding to each of the sensors based on the global clock and the acquisition mode information of each of the sensors; The following operations are performed for each of the sensors: at the start time of data collection corresponding to the sensor, an acquisition trigger signal corresponding to the sensor is sent to the sensor, so that the sensor starts to collect data of the target to be identified based on the acquisition trigger signal, so as to obtain multiple synchronous sampling data under the same time axis.

4. The target recognition method according to claim 3, characterized in that: The acquisition mode information includes an acquisition frequency; and generating an acquisition trigger signal corresponding to each sensor based on the global clock and the acquisition mode information of each sensor includes: generating a synchronization signal based on the crystal oscillator frequency of the global clock; The synchronization signal is frequency-divided by different multiples to generate an acquisition trigger signal corresponding to each of the sensors; wherein the frequency of the acquisition trigger signal corresponding to each of the sensors is consistent with the acquisition frequency of the sensor.

5. The target recognition method according to any one of claims 1 to 4, characterized in that: The target recognition system includes a fixed field terminal for information exchange with a mobile terminal to be identified; the sensor includes: at least one of a camera, a laser radar and a GNSS antenna installed at a predetermined position of the fixed field terminal.

6. The target recognition method according to claim 1, characterized in that: The obtaining of a starting scanning angle of each sensor based on the reference position coordinates and the acquisition mode information of each sensor includes: Converting the reference position coordinates into the coordinate system of each sensor to obtain reference coordinates corresponding to each sensor; For each of the sensors, the following operations are performed: based on the reference coordinates corresponding to the sensor and the acquisition mode information of the sensor, a starting scanning angle of the sensor is obtained.

7. The target recognition method according to claim 4, characterized in that: The method further comprises: Testing the multiple synchronous sampling data under the same time axis; When the test fails, the acquisition synchronization parameter is updated.

8. The target recognition method according to claim 7, characterized in that: The checking of the multiple synchronous sampling data under the same time axis includes: For each synchronous sampling data, perform the following operations: Calculating the time corresponding to the frame start position of the synchronous sampling data under the preset acquisition synchronization parameters based on the acquisition frequency of the sensor corresponding to the synchronous sampling data; Determining whether the time corresponding to the frame start position of the synchronous sampling data matches the acquisition trigger signal corresponding to the sensor; When there is a match, it is determined that the synchronous sampling data has passed the inspection.

9. The target recognition method according to claim 7, characterized in that: The preset acquisition synchronization parameters include: at least one preset moment, and the synchronous scanning direction of all the sensors corresponding to each preset moment; the checking of the multiple synchronous sampling data under the same time axis includes: For each synchronous sampling data, perform the following operations: Acquire the moment when the sensor corresponding to the synchronous sampling data scans to the synchronous scanning direction as the actual moment; Calculating the difference between the actual time and the preset time; Determining whether the difference is less than a preset deviation threshold; When the difference is smaller than the preset deviation threshold, it is determined that the synchronous sampling data has passed the inspection.

10. The target recognition method according to claim 1, characterized in that: The target recognition system includes: a fixed field terminal that interacts with a mobile terminal to be recognized; the method further includes: Time synchronization is performed on the mobile terminal to be identified and the fixed field terminal.

11. The target recognition method according to claim 10, characterized in that: The time used by the fixed terminal includes: a preset first system time and a preset first local time; the time used by the mobile terminal to be identified includes: a preset second system time and a preset second local time; and the time synchronization of the mobile terminal to be identified and the fixed terminal includes: Obtaining a time difference between the first system time and the second system time as a first time difference; Obtaining a time difference between the first system time and the first local time as a second time difference; Obtaining a time difference between the second system time and the second local time as a third time difference; calculating a time difference between the first local time and the second local time based on the first time difference, the second time difference, and the third time difference; Based on the time difference between the first local time and the second local time, time synchronization is performed on the mobile terminal to be identified and the fixed field terminal.

12. The target recognition method according to claim 11, characterized in that: The time synchronization of the mobile terminal to be identified and the fixed field terminal further includes: detecting a link delay between the fixed field terminal and the mobile terminal to be identified; Determining whether the first time difference is credible based on the link delay and the first time difference; The acquiring a time difference between the first system time and the first local time as a second time difference includes: When the first time difference is credible, the time difference between the first system time and the first local time is obtained as the second time difference.

13. A target recognition system, characterized in that: The method comprises a plurality of sensors, a processor and a storage device; the sensors are used to collect data of the target to be identified; the storage device is suitable for storing a plurality of program codes, and the program codes are suitable for being loaded and run by the processor to execute the target identification method according to any one of claims 1 to 12.

14. A computer-readable storage medium storing a plurality of program codes, characterized in that: The program code is suitable for being loaded and run by a processor to execute the target recognition method according to any one of claims 1 to 12.

Images