WO2024021541A1 - 目标跟踪方法、装置、设备及介质 - Google Patents

目标跟踪方法、装置、设备及介质 Download PDF

Info

Publication number
WO2024021541A1
WO2024021541A1 PCT/CN2023/072289 CN2023072289W WO2024021541A1 WO 2024021541 A1 WO2024021541 A1 WO 2024021541A1 CN 2023072289 W CN2023072289 W CN 2023072289W WO 2024021541 A1 WO2024021541 A1 WO 2024021541A1
Authority
WO
WIPO (PCT)
Prior art keywords
track
target
data set
estimate
current frame
Prior art date
Application number
PCT/CN2023/072289
Other languages
English (en)
French (fr)
Inventor
孙靖虎
刘加欢
吴健
黄力
Original Assignee
惠州市德赛西威智能交通技术研究院有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 惠州市德赛西威智能交通技术研究院有限公司 filed Critical 惠州市德赛西威智能交通技术研究院有限公司
Publication of WO2024021541A1 publication Critical patent/WO2024021541A1/zh

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/66—Radar-tracking systems; Analogous systems
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section

Definitions

  • This application relates to the field of radar technology, such as target tracking methods, devices, equipment and media.
  • This application provides a target tracking method, device, equipment and medium to realize target tracking.
  • a target tracking method including:
  • a target track set corresponding to the tracking target is determined.
  • a target tracking device including:
  • the correlation result determination module is configured to obtain the measurement point data set of the current frame collected by the millimeter wave radar, and obtain the correlation result based on the predetermined prediction result set and the measurement point data set;
  • An update module configured to update the track estimate of the current frame according to the correlation result
  • a track set determination module is configured to determine a target track set corresponding to the tracking target based on the track estimate.
  • an electronic device includes:
  • the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor, so that the at least one processor can execute the method described in any embodiment of the present application.
  • Target tracking methods, devices, equipment and media methods are examples of the present application.
  • a computer-readable storage medium stores computer instructions.
  • the computer instructions are configured to enable the processor to implement any embodiment of the present application when executed. target tracking method.
  • Figure 1 is a flow chart of a target tracking method provided according to Embodiment 1 of the present application.
  • Figures 2a-2b are example diagrams of a target track set according to a target tracking method provided in Embodiment 1 of the present application;
  • Figure 3 is a schematic structural diagram of a target tracking device provided according to Embodiment 3 of the present application.
  • FIG. 4 is a schematic structural diagram of an electronic device that implements the target tracking method according to the embodiment of the present application.
  • Figure 1 is a flow chart of a target tracking method provided by Embodiment 1 of the present application. This embodiment can be applied to target tracking in traffic scenarios.
  • the method can be executed by a target tracking device, and the target tracking device can use hardware. And/or implemented in the form of software, the target tracking device can be configured in an electronic device. As shown in Figure 1, the method includes:
  • millimeter wave radar can be understood as a device that sends electromagnetic waves.
  • the measurement point data set can be understood as the data information corresponding to the moving targets within the measurement range.
  • the prediction result set can be understood as the prediction result of the current frame obtained by predicting the previous frame.
  • the correlation results can be understood as being used to determine whether the prediction result set is related to the data in the measurement point data set.
  • the millimeter-wave radar emits electromagnetic waves, and the echo and the emitted electromagnetic waves are processed to obtain the measurement point data set of the current frame collected by the millimeter-wave radar.
  • the nearest neighbor algorithm can be used for correlation judgment.
  • the nearest neighbor algorithm can be used to calculate the distance between all measurement point data in the current frame and the prediction result, and then select the measurement point data within the radar detection range that is closest to the prediction result.
  • the correlation result is determined to be correlation; if no measurement point data falls within the radar detection range, the correlation result is considered to be non-correlation.
  • the track estimate can be understood as the estimate used to determine the track.
  • the corresponding prediction result can be used to replace the empty measurement point data corresponding to the measurement point, and mark it as pseudo-new trajectory data.
  • pseudo-new track data can be understood as track data that is replaced by prediction results when the target is at a certain measurement point and the millimeter wave radar cannot obtain the measurement point data.
  • the target track set can be understood as a set corresponding to the tracks of the tracked target.
  • the track estimate may include track estimates that do not meet the conditions. If the track estimate is not on the road, the wrong track estimate will be deleted, and the first frame to the current frame can be obtained. The correct track estimate is determined as the target track set corresponding to the tracking target.
  • the first embodiment provides a target tracking method that obtains a target track set by inputting the measurement point data set of the current frame collected by millimeter wave radar and the prediction result of the previous frame into a set association and update algorithm. It realizes the determination of the tracking target track, solves the problem of track breakage during the tracking process, and improves the accuracy of the track results.
  • the step of determining the prediction result set includes:
  • the tracking target can be understood as a target with quantitative measurement point data detected by the millimeter wave radar, such as a vehicle traveling on the road.
  • the target track status data can be understood as dividing the target track into different states and the data corresponding to the different states. It can include new track data, death track data, pseudo-new track data, and stopped track data.
  • the new track data can be understood as the track data corresponding to the sudden appearance of a track, and its corresponding state is the new state
  • the death track data can be understood as the track data corresponding to the breakage of the track in the middle of a track.
  • Trace data, its corresponding status is Death state
  • stopped track data can be understood as track data corresponding to the target stopping at a certain measurement point for a period of time, and its corresponding state is the stopped state.
  • the target track set can be understood as a set of track data corresponding to the tracks that the target has traveled.
  • the target track status data and target track set corresponding to the tracking target in the previous frame can be obtained, and the death track data set in the target track status data and the pseudo-rebirth track data set can be track-associated to obtain
  • the correlation result between the newborn track set and the death track set is judged based on the correlation result to determine whether the status can be updated.
  • the prediction result set corresponding to the current frame is determined, including:
  • different target track states correspond to different target track state data.
  • the pseudo-reborn track data set corresponding to the pseudo-reborn state of the last frame corresponding to the tracking target and the dead target can be obtained.
  • the corresponding death track data subset merge the death data subsets from the first frame to the previous frame corresponding to the tracking target to obtain a multi-frame death track data set.
  • the pseudo-new track data set and the death track data set can be input into the corresponding algorithm for calculation, and the new track whose starting state falls within the gate of the death track data set of the current frame can be obtained.
  • Data set use the new track data set as the possible stop track data set, calculate the interconnected Gaussian likelihood function set of the two based on the possible stop track data set and the death track data set, that is, obtain the correlation between the two Strength, according to the set of interconnected likelihood functions, the track correlation result, that is, the correlation probability, can be calculated through Formula 1:
  • t represents the possible stop track state
  • l represents the death track state
  • ⁇ tl represents the track correlation result of the possible stop track data set and the death track data set
  • G tl represents the interconnected Gaussian likelihood function
  • m represents The number of data in the possible stop track data set
  • T represents the number of data in the death track data set
  • B represents the constant related to clutter density.
  • the death state corresponding to the death track data set is regarded as the stopped state
  • the pseudo-rebirth state corresponding to the pseudo-rebirth track data set is regarded as the rebirth state.
  • the first set threshold can be understood as a value set according to the degree of correlation.
  • the first setting threshold is represented by the symbol ⁇ .
  • max ⁇ tl ⁇ is greater than ⁇ , it can be considered that the degree of correlation between the possible stop track data set and the death track data set is very high. Then the broken part of the track corresponding to the death state can be supplemented by the possible stop track data set, the death state corresponding to the death track data set is updated to the stopped state, and the pseudo-rebirth state corresponding to the pseudo-rebirth track data set is not associated is the pseudo-reborn state corresponding to the reborn track data set as the reborn state.
  • the model and traffic scene need to be modeled first.
  • the model can be modeled under the framework of Bayesian filtering:
  • f k is the state transition function
  • n x ,n v represents the dimension of the vector of state and process noise
  • x k-1 represents the target track status data in the previous frame
  • x R represents the road restriction information in the x direction in the road scene
  • y R represents the road restriction information in the y direction in the road scene.
  • f k is the state transition function, which can be described by different models. Commonly used models include uniform motion model (CV), uniform acceleration motion model (CA), coordinated turning model (CT), etc.
  • CV uniform motion model
  • CA uniform acceleration motion model
  • CT coordinated turning model
  • k-1 is the track prediction value.
  • the airspace-based algorithm calculates the interconnection probability between the track data sets, (4) an interconnection between the newborn track data set and the death track set can be obtained probability matrix.
  • the correlation probability matrix can be used to calculate the probability that each new track originates from a death track, and then determine whether the track corresponding to the track data set should be regarded as an independent track individual or whether the terminated track is in During the movement, it is in a very low speed or even stopped state, achieving accurate judgment of the target state, and obtaining prediction results based on the track data sets in different states.
  • updating the track estimate of the current frame according to the correlation result includes:
  • the first track estimate can be understood as a track estimate that combines measurement point data and prediction results.
  • the measurement points without measurement point data can be replaced using the prediction results, that is, the broken tracks can be supplemented according to the prediction results to obtain a complete first flight path. track estimation values, and convert uncorrelated measurement points into pseudo-new track data under the frame.
  • determining the first track estimate and updating the pseudo-nascent track data set of the current frame includes:
  • the correlation result is that the prediction result value in the prediction result set is correlated with the measurement point data in the measurement point data set, the measurement point data is used as the first track estimate value.
  • the distance between all measurement point data of the current frame and the prediction result is calculated according to the nearest neighbor algorithm, and the correlation result of the measurement point data closest to the prediction result value within the radar detection range is determined as the correlation.
  • the prediction result When the prediction result When the value is associated with the measurement point data, it can be considered that the measurement point data of the measurement point can be obtained through the radar, then the measurement point data of the measurement point is used as the first track estimate value.
  • the correlation result is considered to be non-correlation, and it can be considered that the quantity cannot be obtained through millimeter wave radar.
  • the measurement point data of the measurement point If the track is broken and the measurement point data cannot be obtained at the broken track, the prediction result value can be used instead of the measurement point data of the measurement point, that is, the prediction result value is used as the third If a track estimate value is obtained, the measurement point data is updated to the pseudo-new track data set of the current frame.
  • the first track estimate is input into a preset update formula.
  • a CKF filter can be used.
  • the first track estimate can be input into the CKF filter to process the nonlinear prediction results to obtain a linear prediction result and project it to the measurement prediction of the measurement latitude.
  • the preset update formula the covariance matrix of the filter gain and estimation error under the current frame is calculated, Obtain the track estimate of the current frame while ensuring that the covariance matrix of the estimation error is minimized, as shown in the formula:
  • z k represents the measurement point data obtained in frame k
  • K k represents the filtering gain of frame k.
  • the prediction results and measurement point data are input into the preset update formula and processed through the filter, thereby solving the nonlinear problem of some data and eliminating the influence of errors. Get more accurate track estimates.
  • determining the target track set corresponding to the tracking target based on the track estimate value includes:
  • At least one different track estimate set can be determined, and each track estimate set corresponds to a track.
  • the accuracy of the track estimate is judged in conjunction with the pre-established road model.
  • the track estimate does not belong to the road model, it will not To meet the conditions. It can also be judged based on the maintenance time of the track corresponding to the track estimate value set, where the maintenance time can correspond to the number of frames included in the track estimate value set. When the number of frames is shorter, the maintenance time is shorter. The conditions are not met.
  • track estimate values that do not meet the conditions in each track estimate value set are deleted to obtain a second track estimate value set, including:
  • the road model can be understood as a model established based on road restriction information in each direction.
  • the road model includes road information, and by comparing the road information with the track estimate, it can be determined whether the track estimate is included in the road information. If the track estimate is not included in the road information, it can be considered that the track estimate exceeds the range of the road, then the track estimate that does not belong to the road model is deleted, and a path estimate containing only the correct track estimate is obtained. Set of intermediate track estimates.
  • each frame can be understood as corresponding to this moment, and the number of frames corresponding to multiple consecutive frames can be understood as multiple consecutive moments, and the time period can be determined based on the number of frames.
  • the start frame and the end frame in the intermediate track estimate value set are obtained, and the number of frames corresponding to the intermediate track estimate value set is obtained based on the start frame and the end frame.
  • a certain intermediate track estimate value set includes consecutive If there are five frames of intermediate track estimation values, then the number of frames corresponds to five, and the time interval between every two frames is 50ms. It can be considered that the time period for which the intermediate track estimation value set is maintained is 250ms.
  • the second set threshold can be understood as the frame value corresponding to the set time period.
  • the intermediate track estimate value set may be due to caused by errors.
  • the conventional tracking algorithm may obtain two tracks, but the two tracks belong to the same target, so the two tracks need to be merged.
  • the corresponding associated track data set is found based on the new track data set, the stopped track data set and each second track estimate value set, and the tracks of the starting frame of the two associated track data sets are Data and termination frame track data associate the two track data sets, replace the original two track data sets with the associated target track set, and delete the two associated tracks in the corresponding data set data set.
  • a correlation algorithm can be used to determine the target track set corresponding to the tracking target.
  • the track data set in the following statements is replaced by track.
  • each track data set can be judged based on the above determination method of the correlation result, and the two associated track data sets ⁇ T bj , T di ⁇ are proposed.
  • the correlation algorithm is as follows:
  • dt is the minimum sampling period
  • t k ⁇ (t ei ,t bj ) is the time corresponding to x k .
  • the third embodiment of this embodiment uses the road model to eliminate erroneous track estimates and track estimate sets, and obtains a correct second track estimate set, which improves the accuracy of the algorithm.
  • the obtained second track estimation value set is combined with the actual scene and algorithm principles.
  • Figures 2a-2b are example diagrams of a target track set of a target tracking method provided in Embodiment 1 of the present application.
  • Figure 2a is a target track set tracked using a conventional algorithm
  • Figure 2b is a target track set tracked using the method provided by this application.
  • the two targets may have stopped for a period of time near the y-axis close to the value 200 before driving.
  • the conventional algorithm considers that the tracks before and after a target stops are in a death state, so these two tracks are identified To belong to four targets, the track is broken.
  • the method provided in this embodiment is used to determine the death track data sets of two corresponding death states before and after the stopping time. If it is determined to be in a stopped state, and a correlation judgment is made and it is found that the two are related, then the start frame and the end frame can be found, and the two tracks are deemed to belong to the same target, and they will be merged, so this application is applicable
  • the method provided by the embodiment determines that there are only two targets in this scenario, and solves the problem of track breakage.
  • FIG 3 is a schematic structural diagram of a target tracking device provided in Embodiment 2 of the present application. As shown in Figure 3, the device includes: an association result determination module 31, an update module 32, and a track set determination module 33.
  • the correlation result determination module 31 is configured to obtain the measurement point data set of the current frame collected by the millimeter wave radar, and obtain the correlation result based on the predetermined prediction result set and the measurement point data set;
  • the update module 32 is configured to update the track estimate of the current frame according to the correlation result
  • the track set determination module 33 is configured to determine the target track set corresponding to the tracking target based on the track estimate value.
  • the second embodiment provides a target tracking device that obtains a target track set by inputting the measurement point data set of the current frame collected by millimeter wave radar and the prediction result of the previous frame into a set association and update algorithm. It realizes the determination of the tracking target track, solves the problem of track breakage during the tracking process, and improves the accuracy of the track results.
  • the steps for determining the prediction result set in the correlation result determination module 31 include:
  • the first determination unit is configured to determine the prediction result set corresponding to the current frame based on the target track status data and target track set corresponding to the tracking target in the previous frame.
  • the result set determination unit is set to:
  • the death state corresponding to the death track data set is regarded as the stopped state, and the pseudo-reborn state corresponding to the pseudo-reborn track data set is regarded as the reborn state;
  • update module 32 also includes:
  • the first update unit is configured to determine the first track estimate and update the pseudo-new track data set of the current frame based on the correlation result;
  • the second update unit is configured to input the first track estimate value into a preset update formula to update the track estimate value of the current frame.
  • the first update unit is set to:
  • the correlation result is that the prediction result value in the prediction result set is correlated with the measurement point data in the measurement point data set, the measurement point data is used as the first track estimate;
  • the prediction result value is used as the first track estimate value, and the measurement point data is updated as the pseudo-new track data set of the current frame.
  • the track set determination module 33 also includes:
  • the second determination unit is configured to determine at least one track estimate value set based on the track estimate value
  • the third determination unit is configured to delete the track estimate values that do not meet the conditions in each track estimate value set, and obtain the second track estimate value set;
  • the fourth determination unit is used to determine the target track set corresponding to the tracking target based on the new track data set, the stopped track data set and each second track estimate value set.
  • the third determination unit is set to:
  • the target tracking device provided by the embodiments of this application can execute the target tracking method provided by any embodiment of this application, and has corresponding functional modules and effects for executing the method.
  • FIG. 4 shows a schematic structural diagram of an electronic device 10 that can be used to implement embodiments of the present application.
  • Electronic devices are intended to refer to various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers.
  • Electronic devices may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (eg, helmets, glasses, watches, etc.), and other similar computing devices.
  • the components shown herein, their connections and relationships, and their functions are examples only and are not intended to limit the implementation of the present application as described and/or claimed herein.
  • the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (Read-Only Memory, ROM) 12, a random access memory (Random Access Memory, RAM) 13, etc., wherein the memory stores computer programs that can be executed by at least one processor, and the processor 11 can execute 12 or a computer program loaded from the storage unit 18 into the random access memory (RAM) 13 to perform various appropriate actions and processes.
  • RAM 13 random access memory
  • various programs and data required for the operation of the electronic device 10 can also be stored.
  • the processor 11, the ROM 12 and the RAM 13 are connected to each other via the bus 14.
  • An input/output (I/O) interface 15 is also connected to the bus 14 .
  • the I/O interface 15 Multiple components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc. etc.; and communication unit 19, such as network card, modem, wireless communication transceiver, etc.
  • the communication unit 19 allows the electronic device 10 to exchange information/data with other devices through computer networks such as the Internet and/or various telecommunications networks.
  • Processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU), various dedicated artificial intelligence (Artificial Intelligence, AI) computing chips, various running Machine learning model algorithm processor, digital signal processor (Digital Signal Process, DSP), and any appropriate processor, controller, microcontroller, etc.
  • the processor 11 performs various methods and processes described above, such as the target tracking method.
  • the target tracking method may be implemented as a computer program, which is tangibly embodied in a computer-readable storage medium, such as the storage unit 18 .
  • part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19.
  • the processor 11 may be configured to perform the target tracking method in any other suitable manner (eg, by means of firmware).
  • FPGAs Field Programmable Gate Arrays
  • ASICs Application Specific Integrated Circuits
  • ASSP Application Specific Standard Parts
  • SOC System on Chip
  • CPLD Complex Programmable logic device
  • These various embodiments may include implementation in one or more computer programs executable and/or interpreted on a programmable system including at least one programmable processor, the programmable processor
  • the processor which may be a special purpose or general purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
  • An output device may be a special purpose or general purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
  • An output device may be a special purpose or general purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
  • Computer programs for implementing the methods of the present application may be written in any combination of at least one programming language. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that the computer program, when executed by the processor, causes the functions/operations specified in the flowcharts and/or block diagrams to be implemented.
  • a computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
  • a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device.
  • Computer-readable storage media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any suitable combination of the foregoing.
  • the computer-readable storage medium may be a machine-readable signal medium.
  • machine-readable storage media examples include at least one wire-based electrical connection, a portable computer disk, a hard disk, random access memory (RAM), read only memory (ROM), Erasable Programmable Read -Only Memory, EPROM) or flash memory, optical fiber, portable compact disk read-only memory (Compact Disc Read-Only Memory, CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above.
  • RAM random access memory
  • ROM read only memory
  • EPROM Erasable Programmable Read -Only Memory
  • flash memory optical fiber
  • portable compact disk read-only memory Compact Disc Read-Only Memory
  • CD-ROM Compact Disc Read-Only Memory
  • magnetic storage device or any suitable combination of the above.
  • the systems and techniques described herein may be implemented on an electronic device having a display device (eg, a cathode ray tube (CRT) or a liquid crystal) for displaying information to the user. (Liquid Crystal Display, LCD monitor); and a keyboard and pointing device (such as a mouse or trackball) through which the user can configured to provide input to electronic devices.
  • a display device eg, a cathode ray tube (CRT) or a liquid crystal
  • LCD monitor Liquid Crystal Display, LCD monitor
  • keyboard and pointing device such as a mouse or trackball
  • Other kinds of devices may also be configured to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and may be provided in any form, including Acoustic input, voice input or tactile input) to receive input from the user.
  • the systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes a first component (e.g., an application server), or a computing system that includes front-end components (e.g., as a data server) , a user's computer having a graphical user interface or web browser through which the user can interact with implementations of the systems and techniques described herein), or including such backend components, the first components, or any combination of front-end components in a computing system.
  • the components of the system may be interconnected by any form or medium of digital data communication (eg, a communications network). Examples of communication networks include: Local Area Network (LAN), Wide Area Network (WAN), blockchain network, and the Internet.
  • Computing systems may include clients and servers.
  • Clients and servers are generally remote from each other and typically interact over a communications network.
  • the relationship of client and server is created by computer programs running on corresponding computers and having a client-server relationship with each other.
  • the server can be a cloud server, also known as cloud computing server or cloud host. It is a host product in the cloud computing service system to solve the problems that exist in traditional physical host and virtual private server (VPS) services. It has the disadvantages of difficult management and weak business scalability.
  • VPN virtual private server

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

一种目标跟踪方法、装置、设备及介质，其中，目标跟踪方法包括：获取毫米波雷达采集的当前帧的量测点数据集，根据预确定的预测结果集及量测点数据集得到关联结果（S110）；根据关联结果更新当前帧的航迹估计值（S120）；根据航迹估计值，确定跟踪目标对应的目标航迹集（S130）。

Description

目标跟踪方法、装置、设备及介质
本申请要求在2022年7月27日提交中国专利局、申请号为202210892722.7的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请涉及雷达技术领域，例如涉及目标跟踪方法、装置、设备及介质。
背景技术
在现在的交通场景下，可能会出现拥堵等道路情况，导致车辆不能连续正常行驶，车辆的走停走现象越来越普遍。
相关技术在对走停走目标进行跟踪时，当目标的回波被障碍物遮挡，导致无法探测到目标，或者当目标速度减小至雷达测速门限以下时，会使运动目标变为停止目标，则跟踪结果会出现航迹断裂的情况。
发明内容
本申请提供了一种目标跟踪方法、装置、设备及介质，实现了对目标的跟踪。
根据本申请的第一方面，提供了一种目标跟踪方法，包括：
获取毫米波雷达采集的当前帧的量测点数据集，根据预确定的预测结果集及所述量测点数据集得到关联结果；
根据所述关联结果更新当前帧的航迹估计值；
根据所述航迹估计值，确定跟踪目标对应的目标航迹集。
根据本申请的第二方面，提供了一种目标跟踪装置，包括：
关联结果确定模块，设置为获取毫米波雷达采集的当前帧的量测点数据集，根据预确定的预测结果集及所述量测点数据集得到关联结果；
更新模块，设置为根据所述关联结果更新当前帧的航迹估计值；
航迹集确定模块，设置为根据所述航迹估计值，确定跟踪目标对应的目标航迹集。
根据本申请的第三方面，提供了一种电子设备，所述电子设备包括：
至少一个处理器；以及
与所述至少一个处理器通信连接的存储器；其中，
所述存储器存储有可被所述至少一个处理器执行的计算机程序，所述计算机程序被所述至少一个处理器执行，以使所述至少一个处理器能够执行本申请任一实施例所述的目标跟踪方法、装置、设备及介质方法。
根据本申请的另一方面，提供了一种计算机可读存储介质，所述计算机可读存储介质存储有计算机指令，所述计算机指令设置为使处理器执行时实现本申请任一实施例所述的目标跟踪方法。
附图说明
图1是根据本申请实施例一提供的一种目标跟踪方法的流程图；
图2a-2b是根据本申请实施例一提供的一种目标跟踪方法的目标航迹集示例图；
图3是根据本申请实施例三提供的一种目标跟踪装置的结构示意图；
图4是实现本申请实施例的目标跟踪方法的电子设备的结构示意图。
具体实施方式
需要说明的是，本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本申请的实施例能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了 一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
实施例一
图1为本申请实施例一提供了一种目标跟踪方法的流程图，本实施例可适用于交通场景下的目标跟踪情况，该方法可以由目标跟踪装置来执行，该目标跟踪装置可以采用硬件和/或软件的形式实现，该目标跟踪装置可配置于电子设备中。如图1所示，该方法包括：
S110、获取毫米波雷达采集的当前帧的量测点数据集，根据预确定的预测结果集及量测点数据集得到关联结果。
可以知道的是，根据毫米波雷达的特性，当目标速度减小至雷达测速门限以下或可能目标的运动速度与雷达呈现一定角度，使其径向速度降低至雷达测速门限以下时，目标的回波落入零多普勒频率附近，或者目标的回波被障碍物遮挡，均会导致无法探测到目标，则无法通过毫米波雷达获取该目标对应的量测点数据，则需要对上一帧包括的数据信息进行预测，得到预测结果集。
在本实施例中，毫米波雷达可以理解为发送电磁波的器件。量测点数据集可以理解为量测范围内的移动目标对应的数据信息。预测结果集可以理解为对上一帧进行预测获得的当前帧的预测结果。关联结果可以理解为用于判断预测结果集与量测点数据集中的数据是否相关。
示例性的，毫米波雷达发出电磁波，将回波与发出的电磁波进行处理可以获取毫米波雷达采集的当前帧的量测点数据集，根据量测点数据集中各量测点数据与预确定的预测结果集中的预测结果进行关联判断。可选地，可以采用最近邻域算法进行关联判断，通过最近邻域算法计算当前帧所有量测点数据与预测结果之间的距离，将雷达探测范围内与预测结果距离最近的量测点数据的关联结果确定为关联；若没有量测点数据落在雷达探测范围内，则认为关联结果为非关联。
S120、根据关联结果更新当前帧的航迹估计值。
在本实施例中，航迹估计值可以理解为用于确定航迹的估计值。
示例性的，可以根据关联结果，当量测点没有量测点数据时可以用对应的预测结果替代该量测点对应的空的量测点数据，并将其标记为伪新生航迹数据。其中，伪新生航迹数据可以理解为目标在某一量测点时，毫米波雷达无法获取量测点数据，则利用预测结果代替的航迹数据。将关联的量测点数据与使用预测结果替代的量测点数据带入预先设定的更新公式，可以获得更新后的当前帧的航迹估计值。
S130、根据航迹估计值，确定跟踪目标对应的目标航迹集。
在本实施例中，目标航迹集可以理解为跟踪目标行驶的航迹对应的集合。
示例性的，航迹估计值中可能包括不满足条件的航迹估计值，如航迹估计值不在该条道路中等情况，则删除错误的航迹估计值，可以将第一帧至当前帧获取的正确的航迹估计值确定为该跟踪目标对应的目标航迹集。
本实施例一提供的一种目标跟踪方法，通过将毫米波雷达采集的当前帧的量测点数据集与上一帧的预测结果输入设定的关联及更新算法中，获得目标航迹集，实现了对跟踪目标航迹的确定，解决了跟踪过程中航迹断裂的问题，提高了航迹结果的准确度。
作为本实施例的第一可选实施例，预测结果集的确定步骤包括：
a1、根据上一帧跟踪目标对应的目标航迹状态数据及目标航迹集，确定当前帧对应的预测结果集。
在本实施例中，跟踪目标可以理解为毫米波雷达探测到的有量测点数据的目标，如：道路中行驶的车辆等。目标航迹状态数据可以理解为将目标航迹分为不同的状态，不同状态所对应的数据。可以包括新生航迹数据、死亡航迹数据、伪新生航迹数据、停止航迹数据。其中，新生航迹数据可以理解为突然出现了一条航迹对应的航迹数据，其对应的状态为新生状态；死亡航迹数据可以理解为在一条航迹中间出现了航迹断裂情况对应的航迹数据，其对应的状态为 死亡状态；停止航迹数据可以理解为目标在某一量测点处停止了一段时间对应的航迹数据，其对应的状态为停止状态。
在本实施例中，目标航迹集可以理解为目标已经走过的航迹对应的航迹数据的集合。
示例性的，可以获取上一帧跟踪目标对应的目标航迹状态数据及目标航迹集，将目标航迹状态数据中的死亡航迹数据集与伪新生航迹数据集进行航迹关联，得到新生航迹集与死亡航迹集的关联结果，根据关联结果判断，状态是否可以更新。
示例性的，根据上一帧跟踪目标对应的目标航迹状态数据及目标航迹集，确定当前帧对应的预测结果集，包括：
a11、获取上一帧伪新生状态对应的伪新生航迹数据集以及死亡状态对应的死亡航迹数据子集，并将第一帧至上一帧的死亡航迹数据子集合并为死亡航迹数据集。
示例性的，不同目标航迹状态均对应着不同的目标航迹状态数据，从目标航迹状态数据中可以获取跟踪目标对应的上一帧伪新生状态对应的伪新生航迹数据集以及死亡目标对应的死亡航迹数据子集，将该跟踪目标对应的第一帧至上一帧的死亡数据子集进行合并，得到多帧下的死亡航迹数据集。
a12、根据伪新生航迹数据集及死亡航迹数据集，确定航迹关联结果。
示例性的，可以将伪新生航迹数据集与死亡航迹数据集输入相应的算法中进行计算，获得航迹起始状态落入当前帧死亡航迹数据集的波门之内的新生航迹数据集，将该新生航迹数据集作为可能停止航迹数据集，根据可能停止航迹数据集与死亡航迹数据集计算两者的互联高斯似然函数集，即得到两者之间的关联强度，根据互联似然函数集可以通过公式1计算航迹关联结果，即关联概率：
式中，t表示可能停止航迹状态，l表示死亡航迹状态，βtl表示可能停止航迹数据集与死亡航迹数据集的航迹关联结果，Gtl表示互联高斯似然函数，m表示可能停止航迹数据集中的数据个数，T表示死亡航迹数据集中的数据个数，B表示有关杂波密度的常数。
a13、当航迹关联结果大于第一设定阈值时，将死亡航迹数据集对应的死亡状态作为停止状态，将伪新生航迹数据集对应的伪新生状态作为新生状态。
在本实施例中，第一设定阈值可以理解根据关联程度设定的数值。
示例性的，在本实施例中将第一设定阈值以符号β表示，当max{βtl}大于β时，可以认为可能停止航迹数据集与死亡航迹数据集的关联程度很高，则死亡状态对应的航迹断裂部分可以由可能停止航迹数据集进行补充，将死亡航迹数据集对应的死亡状态更新为停止状态，将伪新生航迹数据集对应的伪新生状态中未关联的为新生航迹数据集对应的伪新生状态作为新生状态。
a14、将新生状态对应的新生航迹数据集、停止状态对应的停止航迹数据集以及目标航迹集输入预先设定的预测公式，确定当前帧对应的预测结果集。
示例性的，在预测前首先需要对模型及交通场景进行建模。可选地，可以采用贝叶斯滤波的框架下对模型建模：
xk＝fk(xk-1,vk-1)    (2)
其中：fk为状态转移函数，nx,nv表示状态与过程噪声的向量的维数，是关于xk-1所有可能的非线性函数的空间集合，表示过程噪声序列，xk-1表示在上一帧的目标航迹状态数据，表示自然数集。
对于交通场景进行建模得到道路模型，则场景中的道路信息为Mt：
其中，xR表示道路场景中x方向的道路限制信息，yR表示道路场景中y方向的道路限制信息。fk为状态转移函数，可以采用不同的模型来描述，常用的模型有匀速运动模型(CV)，匀加速运动模型(CA)，协同转弯模型(CT)等，对于CA模型来说，公式(3)转化为：
xk|k-1＝fkxk-1+vk-1    (4)
式中，xk|k-1为航迹预测值。
本实施例的第一实施例通过这样的设置，基于空域的算法对航迹数据集之间的互联概率进行计算，(4)可以获得一个关于新生航迹数据集和死亡航迹集间的互联概率矩阵。通过该关联概率矩阵可计算每一条新生航迹源于死亡航迹的概率大小，进而判断该航迹数据集对应的航迹是否应该视为一个独立航迹个体或是判断已终止航迹是否在运动过程中处于一段极低速甚至停止的状态中，实现了对目标状态的准确判断，并根据不同状态下的航迹数据集获得预测结果。
作为本实施例的第二可选实施例，在上述实施例的基础上，根据关联结果更新当前帧的航迹估计值，包括：
a2、根据关联结果，确定第一航迹估计值并更新当前帧的伪新生航迹数据集。
在本实施例中，第一航迹估计值可以理解为结合了量测点数据与预测结果的航迹估计值。
示例性的，根据关联结果，可以没有量测点数据的量测点利用预测结果进行更换，即可以认为将断裂的航迹根据预测结果进行补充，得到完整的第一航 迹估计值，并将没有关联的量测点转化为该帧下的伪新生航迹数据。
示例性的，根据关联结果，确定第一航迹估计值并更新当前帧的伪新生航迹数据集，包括：
a21、当关联结果为预测结果集中的预测结果值与量测点数据集中的量测点数据关联时，将量测点数据作为第一航迹估计值。
示例性的，根据最近邻域算法计算出当前帧所有量测点数据与预测结果的距离，将雷达探测范围内与预测结果值距离最近的量测点数据的关联结果确定为关联，当预测结果值与量测点数据关联时，可以认为通过雷达可以获取该量测点的量测点数据，则该量测点以量测点数据作为第一航迹估计值。
a22、否则，将预测结果值作为第一航迹估计值，将量测点数据更新为当前帧的伪新生航迹数据集。
示例性的，当没有量测点数据落在毫米波雷达探测范围内，则预测结果值周围没有对应的量测点数据，则认为关联结果为非关联，可以认为无法通过毫米波雷达获取该量测点的量测点数据，如航迹断裂情况下，断裂航迹处无法获取量测点数据，可以将预测结果值代替该量测点的量测点数据，即则将预测结果值作为第一航迹估计值，则将该量测点数据更新为当前帧的伪新生航迹数据集。
b2、将第一航迹估计值输入预先设定的更新公式更新当前帧的航迹估计值。
可以知道的是，由于经过CA模型处理的预测结果是非线性的，在某些时间段中可能不满足线性的要求，则会导致预测结果的不准确，因此可以考虑使用滤波器对包含量测点数据及预测结果的第一航迹估计值进行更新解决非线性问题。
示例性的，将第一航迹估计值输入预先设定的更新公式。可选地，可以采用CKF滤波器，可以将第一航迹估计值输入CKF滤波器中将其中非线性的预测结果进行处理，得到线性的预测结果及将其投影到量测纬度的量测预测结果，根据预先设定的更新公式，计算当前帧下的滤波增益及估计误差的协方差矩阵， 在保证估计误差的协方差矩阵最小的情况下获得当前帧的航迹估计值，如公式：
其中zk表示k帧得到的量测点数据，Kk表示k帧的滤波增益，表示经过滤波器处理后的对k帧的预测结果，表示对k帧的量测预测结果。
本实施例的第二实施例通过这样的设置，将预测结果及量测点数据输入预先设定的更新公式并通过滤波器的处理，解决了部分数据的非线性问题，消除了误差的影响，得到更加准确的航迹估计值。
作为本实施例的第三可选实施例，根据航迹估计值，确定跟踪目标对应的目标航迹集，包括：
a3、根据航迹估计值，确定至少一个航迹估计值集。
示例性的，将已存在的航迹估计值进行合并，可以确定出至少一个不同的航迹估计值集，每个航迹估计值集对应着一条航迹。
b3、删除各航迹估计值集中不满足条件的航迹估计值，得到第二航迹估计值集。
示例性的，为了保证目标航迹集的准确性，首先需要对航迹估计值的准确性进行判断，可以结合预先建立的道路模型进行判断，当航迹估计值不属于道路模型时，则不满足条件。也可以根据航迹估计值集所对应的航迹的维持时间进行判断，其中，维持时间可以对应着航迹估计值集中所包括的帧数，当帧数较短时，即维持时间较短，则不满足条件。
示例性的，删除各航迹估计值集中不满足条件的航迹估计值，得到第二航迹估计值集，包括：
b31、基于预先建立的道路模型，删除不属于道路模型的航迹估计值，得到中间航迹估计值集。
在本实施例中，道路模型可以理解为根据各方向的道路限制信息建立的模型。
示例性的，道路模型中包括道路信息，根据道路信息与航迹估计值进行比对，可以判断是否航迹估计值是否在包含在道路信息中。如当航迹估计值不包含在道路信息中，可以认为该航迹估计值超出了该条道路的范围，则删除不属于道路模型的航迹估计值，得到只包含正确的航迹估计值的中间航迹估计值集。
b32、确定各中间航迹估计值集中包括的帧数。
可以知道的是，每一帧可以理解为对应这一个时刻，连续多帧对应的帧数可以理解为连续的多个时刻，则可以根据帧数确定时间段。
示例性的，获取中间航迹估计值集中的起始帧与终止帧，根据起始帧及终止帧获得该中间航迹估计值集对应的帧数，如某一条中间航迹估计值集中包括连续五帧中间航迹估计值，则帧数对应为五，每两帧的时间间隔为50ms，则可以认为该中间航迹估计值集维持的时间段为250ms。
b33、删除帧数小于第二设定阈值的中间航迹估计值集，得到第二航迹估计值集。
在本实施例中，第二设定阈值可以理解为与设定时间段相对应的帧数值。
可以知道的是，当帧数过少时对应着该航迹维持的时间段很短，即该中间航迹估计值集对应的航迹的距离过短，则该中间航迹估计值集可能是由于误差导致的。
示例性的，将各中间航迹估计值集对应的帧数与第二设定阈值进行比对，删除小于第二设定阈值的中间航迹估计值集，可以得到维持的时间段满足第二设定阈值的第二航迹估计值集。
c3、根据新生航迹数据集、停止航迹数据集及各第二航迹估计值集，确定跟踪目标对应的目标航迹集。
可以知道的是，一个跟踪目标停止了一段时间后继续行驶，常规跟踪算法可能得到两条航迹，但两条航迹属于同一个目标，则需要对两条航迹进行合并。
示例性的，根据新生航迹数据集、停止航迹数据集及各第二航迹估计值集中找出对应的相关联的航迹数据集，根据两关联航迹数据集的起始帧航迹数据及终止帧航迹数据将两个航迹数据集进行关联，将关联后的目标航迹集替换原有的两条航迹数据集，并在对应的数据集中删除两个已关联的航迹数据集。
可选的，可以采取关联算法确定跟踪目标对应的目标航迹集。为了便于理解将下述陈述中的航迹数据集替换为航迹。假设死亡航迹数据集(所有死亡航迹的集合)为Tdeath＝{Td1,Td2,…,Tdm}，Tdi,i＝1,2,…,m为某一条停止航迹。新生航迹数据集(所有新生航迹的集合)为TB＝{Tb1,Tb2,…,Tbn}，Tbj,j＝1,2,…,n为某一条新生航迹。可以根据上述关联结果的确定方法对各航迹数据集的关联性进行判断，提出关联的两个航迹数据集{Tbj,Tdi}，根据航迹Tei的终止帧航迹数据xdi和航迹Tbj的起始帧航迹数据xbj将两条航迹关联起来。关联算法如下：
对于终止帧航迹数据xdi、终止帧帧数tdi、起始帧航迹数据xbj和起始帧帧数tbj求出其有关于状态的微量：
根据状态微量补充丢失的空白帧：
式中，dt为最小采样周期，tk∈(tei,tbj)为xk对应时刻。
将航迹对{Tbj,Tei}生成的新航迹置入存活航迹集TS中，删除停止航迹集TEND和新生航迹集TB对应航迹。
本实施例的第三实施例通过这样的设置，利用道路模型，排除错误的航迹估计值与航迹估计值集，得到了正确的第二航迹估计值集，提高了算法的精度，根据得到的第二航迹估计值集结合实际场景及算法原理，当跟踪目标在趋近于停止的运动状态时，目标的空间状态值不会发生较大的变化，可以实现对停止时间较长的目标的跟踪。
图2a-2b为本申请实施例一提供的一种目标跟踪方法的目标航迹集示例图。图2a为使用常规算法跟踪的目标航迹集，图2b为使用本申请提供的方法跟踪的目标航迹集。
如图2a所示，两个目标在y轴靠近数值200的附近可能停止了一段时间再进行行驶，则常规算法认为一个目标停止前后的航迹均为死亡状态，所以将这两条航迹认定为属于四个目标，航迹是断裂的。
如图2b所示，经过本申请实施例一提供的一种目标跟踪方法处理后，针对一个跟踪目标经过本实施例提供的方法判定停止时刻前后对应的两个死亡状态下的死亡航迹数据集认定为停止状态，并进行关联判定发现二者是相关联的，则可以通过找到起始帧与终止帧，则认定为这两条航迹属于一个目标，则将其进行合并，所以适用本申请实施例提供的方法认定该场景下只有两个目标，解决了航迹断裂的问题。
实施例二
图3为本申请实施例二提供的一种目标跟踪装置的结构示意图。如图3所示，该装置包括：关联结果确定模块31、更新模块32、航迹集确定模块33。
其中，关联结果确定模块31，设置为获取毫米波雷达采集的当前帧的量测点数据集，根据预确定的预测结果集及量测点数据集得到关联结果；
更新模块32，设置为根据关联结果更新当前帧的航迹估计值；
航迹集确定模块33，设置为根据航迹估计值，确定跟踪目标对应的目标航迹集。
本实施例二提供的一种目标跟踪装置，通过将毫米波雷达采集的当前帧的量测点数据集与上一帧的预测结果输入设定的关联及更新算法中，获得目标航迹集，实现了对跟踪目标航迹的确定，解决了跟踪过程中航迹断裂的问题，提高了航迹结果的准确度。
可选的，关联结果确定模块31中预测结果集的确定步骤，包括：
第一确定单元，设置为根据上一帧跟踪目标对应的目标航迹状态数据及目标航迹集，确定当前帧对应的预测结果集。
其中，结果集确定单元设置为：
获取上一帧伪新生状态对应的伪新生航迹数据集以及死亡目标对应的死亡航迹数据子集，并将第一帧至所述上一帧的死亡航迹数据子集合并为死亡航迹数据集；
根据伪新生航迹数据集及死亡航迹数据集，确定航迹关联结果；
当所述航迹关联结果大于设定阈值时，将死亡航迹数据集对应的死亡状态作为停止状态，将伪新生航迹数据集对应的伪新生状态作为新生状态；
将新生状态对应的新生航迹数据集、停止状态对应的停止航迹数据集以及目标航迹集输入预先设定的预测公式，确定当前帧对应的预测结果集。
可选的，更新模块32，还包括：
第一更新单元，设置为根据关联结果，确定第一航迹估计值并更新当前帧的伪新生航迹数据集；
第二更新单元，设置为将第一航迹估计值输入预先设定的更新公式更新当前帧的航迹估计值。
其中，第一更新单元设置为：
当关联结果为预测结果集中的预测结果值与量测点数据集中的量测点数据关联时，将量测点数据作为第一航迹估计值；
否则，将预测结果值作为第一航迹估计值，将量测点数据更新为当前帧的伪新生航迹数据集。
可选的，航迹集确定模块33，还包括：
第二确定单元，设置为根据航迹估计值，确定至少一个航迹估计值集；
第三确定单元，设置为删除各航迹估计值集中不满足条件的航迹估计值，得到第二航迹估计值集；
第四确定单元，用于根据新生航迹数据集、停止航迹数据集及各第二航迹估计值集，确定跟踪目标对应的目标航迹集。
其中，第三确定单元设置为：
基于预先建立的道路模型，删除不属于道路模型的航迹估计值，得到中间航迹估计值集；
确定各中间航迹估计值集中包括的帧数；
删除帧数小于设定阈值的中间航迹估计值集，得到第二航迹估计值集。
本申请实施例所提供的目标跟踪装置可执行本申请任意实施例所提供的目标跟踪方法，具备执行方法相应的功能模块和效果。
实施例三
图4示出了可以用来实施本申请的实施例的电子设备10的结构示意图。电子设备旨在表示各种形式的数字计算机，诸如，膝上型计算机、台式计算机、工作台、个人数字助理、服务器、刀片式服务器、大型计算机、和其它适合的计算机。电子设备还可以表示各种形式的移动装置，诸如，个人数字处理、蜂窝电话、智能电话、可穿戴设备(如头盔、眼镜、手表等)和其它类似的计算装置。本文所示的部件、它们的连接和关系、以及它们的功能仅仅作为示例，并且不意在限制本文中描述的和/或者要求的本申请的实现。
如图4所示，电子设备10包括至少一个处理器11，以及与至少一个处理器11通信连接的存储器，如只读存储器(Read-Only Memory，ROM)12、随机访问存储器(Random Access Memory，RAM)13等，其中，存储器存储有可被至少一个处理器执行的计算机程序，处理器11可以根据存储在只读存储器(ROM) 12中的计算机程序或者从存储单元18加载到随机访问存储器(RAM)13中的计算机程序，来执行各种适当的动作和处理。在RAM 13中，还可存储电子设备10操作所需的各种程序和数据。处理器11、ROM 12以及RAM 13通过总线14彼此相连。输入/输出(Input/Output，I/O)接口15也连接至总线14。
电子设备10中的多个部件连接至I/O接口15，包括：输入单元16，例如键盘、鼠标等；输出单元17，例如各种类型的显示器、扬声器等；存储单元18，例如磁盘、光盘等；以及通信单元19，例如网卡、调制解调器、无线通信收发机等。通信单元19允许电子设备10通过诸如因特网的计算机网络和/或各种电信网络与其他设备交换信息/数据。
处理器11可以是各种具有处理和计算能力的通用和/或专用处理组件。处理器11的一些示例包括但不限于中央处理单元(Central Processing Unit，CPU)、图形处理单元(Graphics Processing Unit，GPU)、各种专用的人工智能(Artificial Intelligence，AI)计算芯片、各种运行机器学习模型算法的处理器、数字信号处理器(Digital Signal Process，DSP)、以及任何适当的处理器、控制器、微控制器等。处理器11执行上文所描述的各个方法和处理，例如目标跟踪方法。
在一些实施例中，目标跟踪方法可被实现为计算机程序，其被有形地包含于计算机可读存储介质，例如存储单元18。在一些实施例中，计算机程序的部分或者全部可以经由ROM 12和/或通信单元19而被载入和/或安装到电子设备10上。当计算机程序加载到RAM 13并由处理器11执行时，可以执行上文描述的目标跟踪方法的至少一个步骤。备选地，在其他实施例中，处理器11可以通过其他任何适当的方式(例如，借助于固件)而被配置为执行目标跟踪方法。
本文中以上描述的系统和技术的各种实施方式可以在数字电子电路系统、集成电路系统、场可编程门阵列(Field Programmable Gate Array，FPGA)、专用集成电路(Application Specific Integrated Circuit，ASIC)、专用标准产品(Application Specific Standard Parts，ASSP)、系统级芯片(System on Chip，SOC)、复杂可编程逻辑设备(Complex Programmable logic device，CPLD)、计算机硬 件、固件、软件、和/或它们的组合中实现。这些各种实施方式可以包括：实施在一个或者多个计算机程序中，该一个或者多个计算机程序可在包括至少一个可编程处理器的可编程系统上执行和/或解释，该可编程处理器可以是专用或者通用可编程处理器，可以从存储系统、至少一个输入装置、和至少一个输出装置接收数据和指令，并且将数据和指令传输至该存储系统、该至少一个输入装置、和该至少一个输出装置。
用于实施本申请的方法的计算机程序可以采用至少一个编程语言的任何组合来编写。这些计算机程序可以提供给通用计算机、专用计算机或其他可编程数据处理装置的处理器，使得计算机程序当由处理器执行时使流程图和/或框图中所规定的功能/操作被实施。计算机程序可以完全在机器上执行、部分地在机器上执行，作为独立软件包部分地在机器上执行且部分地在远程机器上执行或完全在远程机器或服务器上执行。
在本申请的上下文中，计算机可读存储介质可以是有形的介质，其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的计算机程序。计算机可读存储介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备，或者上述内容的任何合适组合。备选地，计算机可读存储介质可以是机器可读信号介质。机器可读存储介质的示例会包括基于至少一个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(Erasable Programmable Read-Only Memory，EPROM)或快闪存储器、光纤、便捷式紧凑盘只读存储器(Compact Disc Read-Only Memory，CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
为了提供与用户的交互，可以在电子设备上实施此处描述的系统和技术，该电子设备具有：用于向用户显示信息的显示装置(例如，阴极射线管(Cathode Ray Tube，CRT)或者液晶显示器(Liquid Crystal Display，LCD)监视器)；以及键盘和指向装置(例如，鼠标或者轨迹球)，用户可以通过该键盘和该指向装 置来将输入提供给电子设备。其它种类的装置还可以设置为提供与用户的交互；例如，提供给用户的反馈可以是任何形式的传感反馈(例如，视觉反馈、听觉反馈、或者触觉反馈)；并且可以用任何形式(包括声输入、语音输入或者、触觉输入)来接收来自用户的输入。
可以将此处描述的系统和技术实施在包括后台部件的计算系统(例如，作为数据服务器)、或者包括第一件部件的计算系统(例如，应用服务器)、或者包括前端部件的计算系统(例如，具有图形用户界面或者网络浏览器的用户计算机，用户可以通过该图形用户界面或者该网络浏览器来与此处描述的系统和技术的实施方式交互)、或者包括这种后台部件、第一件部件、或者前端部件的任何组合的计算系统中。可以通过任何形式或者介质的数字数据通信(例如，通信网络)来将系统的部件相互连接。通信网络的示例包括：局域网(Local Area Network，LAN)、广域网(Wide Area Network，WAN)、区块链网络和互联网。
计算系统可以包括客户端和服务器。客户端和服务器一般远离彼此并且通常通过通信网络进行交互。通过在相应的计算机上运行并且彼此具有客户端-服务器关系的计算机程序来产生客户端和服务器的关系。服务器可以是云服务器，又称为云计算服务器或云主机，是云计算服务体系中的一项主机产品，以解决了传统物理主机与虚拟专用服务器(Virtual Private Server，VPS)服务中，存在的管理难度大，业务扩展性弱的缺陷。
应该理解，可以使用上面所示的各种形式的流程，重新排序、增加或删除步骤。例如，本申请中记载的各步骤可以并行地执行也可以顺序地执行也可以不同的次序执行，只要能够实现本申请的技术方案所期望的结果，本文在此不进行限制。

Claims (10)

  1. 一种目标跟踪方法，包括：
    获取毫米波雷达采集的当前帧的量测点数据集，根据预确定的预测结果集及所述量测点数据集得到关联结果；
    根据所述关联结果更新当前帧的航迹估计值；
    根据所述航迹估计值，确定跟踪目标对应的目标航迹集。
  2. 根据权利要求1所述的方法，其中，所述预测结果集的确定步骤包括：
    根据上一帧跟踪目标对应的目标航迹状态数据及目标航迹集，确定当前帧对应的预测结果集。
  3. 根据权利要求2所述的方法，其中，所述根据上一帧跟踪目标对应的目标航迹状态数据及目标航迹集，确定当前帧对应的预测结果集，包括：
    获取上一帧伪新生状态对应的伪新生航迹数据集以及死亡状态对应的死亡航迹数据子集，并将第一帧至所述上一帧的死亡航迹数据子集合并为死亡航迹数据集；
    根据所述伪新生航迹数据集及所述死亡航迹数据集，确定航迹关联结果；
    当所述航迹关联结果大于第一设定阈值时，将所述死亡航迹数据集对应的死亡状态作为停止状态，将所述伪新生航迹数据集对应的伪新生状态作为新生状态；
    将所述新生状态对应的新生航迹数据集、停止状态对应的停止航迹数据集以及目标航迹集输入预先设定的预测公式，确定当前帧对应的预测结果集。
  4. 根据权利要求1所述的方法，其中，所述根据所述关联结果更新当前帧的航迹估计值，包括：
    根据所述关联结果，确定第一航迹估计值并更新当前帧的伪新生航迹数据集；
    将第一航迹估计值输入预先设定的更新公式更新当前帧的航迹估计值。
  5. 根据权利要求4所述的方法，其中，所述根据所述关联结果，确定第一航迹估计值并更新当前帧的伪新生航迹数据集，包括：
    当所述关联结果为所述预测结果集中的预测结果值与所述量测点数据集中的量测点数据关联时，将所述量测点数据作为第一航迹估计值；
    否则，将所述预测结果值作为第一航迹估计值，将量测点数据更新为当前帧的伪新生航迹数据集。
  6. 根据权利要求3所述的方法，其中，所述根据所述航迹估计值，确定跟踪目标对应的目标航迹集，包括：
    根据所述航迹估计值，确定至少一个航迹估计值集；
    删除各所述航迹估计值集中不满足条件的航迹估计值，得到第二航迹估计值集；
    根据所述新生航迹数据集、所述停止航迹数据集及各所述第二航迹估计值集，确定所述跟踪目标对应的目标航迹集。
  7. 根据权利要求6所述的方法，其中，所述删除各所述航迹估计值集中不满足条件的航迹估计值，得到第二航迹估计值集，包括：
    基于预先建立的道路模型，删除不属于所述道路模型的航迹估计值，得到中间航迹估计值集；
    确定各所述中间航迹估计值集中包括的帧数；
    删除帧数小于第二设定阈值的中间航迹估计值集，得到所述第二航迹估计值集。
  8. 一种目标跟踪装置，包括：
    关联结果确定模块，设置为获取毫米波雷达采集的当前帧的量测点数据集，根据预确定的预测结果集及所述量测点数据集得到关联结果；
    更新模块，设置为根据所述关联结果更新当前帧的航迹估计值；
    航迹集确定模块，设置为根据所述航迹估计值，确定跟踪目标对应的目标航迹集。
  9. 一种电子设备，包括：
    至少一个处理器；以及
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的计算机程序，所述计算机程序被所述至少一个处理器执行，以使所述至少一个处理器能够执行权利要求1-7中任一项所述的目标跟踪方法。
  10. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机指令，所述计算机指令设置为使处理器执行时实现权利要求1-7中任一项所述的目标跟踪方法。
PCT/CN2023/072289 2022-07-27 2023-01-16 目标跟踪方法、装置、设备及介质 WO2024021541A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210892722.7A CN115755019A (zh) 2022-07-27 2022-07-27 目标跟踪方法、装置、设备及介质
CN202210892722.7 2022-07-27

Publications (1)

Publication Number Publication Date
WO2024021541A1 true WO2024021541A1 (zh) 2024-02-01

Family

ID=85349114

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/072289 WO2024021541A1 (zh) 2022-07-27 2023-01-16 目标跟踪方法、装置、设备及介质

Country Status (2)

Country Link
CN (1) CN115755019A (zh)
WO (1) WO2024021541A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2014211330A (ja) * 2013-04-17 2014-11-13 三菱電機株式会社 目標追尾装置及び目標追尾方法
CN106980114A (zh) * 2017-03-31 2017-07-25 电子科技大学 无源雷达目标跟踪方法
CN110542885A (zh) * 2019-08-13 2019-12-06 北京理工大学 一种复杂交通环境下的毫米波雷达目标跟踪方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2014211330A (ja) * 2013-04-17 2014-11-13 三菱電機株式会社 目標追尾装置及び目標追尾方法
CN106980114A (zh) * 2017-03-31 2017-07-25 电子科技大学 无源雷达目标跟踪方法
CN110542885A (zh) * 2019-08-13 2019-12-06 北京理工大学 一种复杂交通环境下的毫米波雷达目标跟踪方法

Also Published As

Publication number Publication date
CN115755019A (zh) 2023-03-07

Similar Documents

Publication Publication Date Title
CN112560996B (zh) 用户画像识别模型训练方法、设备、可读存储介质及产品
JP2023509572A (ja) 車両を検出するための方法、装置、電子機器、記憶媒体およびコンピュータプログラム
CN110879395B (zh) 障碍物位置预测方法、装置和电子设备
WO2022247915A1 (zh) 融合定位方法、装置、设备、存储介质及程序产品
US11447153B2 (en) Method and apparatus for annotating virtual lane at crossing
EP3943964A2 (en) Method and apparatus for determining positioning information of vehicle, electronic device, storage medium and program product
KR102568948B1 (ko) 장애물 속도를 결정하는 방법, 장치, 전자 기기, 저장 매체 및 프로그램
KR20210137936A (ko) 차량 대기 정보를 결정하는 방법, 전자 기기, 노변 기기, 클라우드 제어 플랫폼, 저장 매체, 및 컴퓨터 프로그램 제품
WO2024051661A1 (zh) 路面不平度的分类识别方法、装置、车辆及存储介质
WO2022160897A1 (zh) 双目视差估计方法、模型训练方法以及相关设备
CN113066302A (zh) 车辆信息预测方法、装置及电子设备
CN115903831A (zh) 一种车辆驾驶控制方法、装置、车辆及存储介质
KR20230008005A (ko) 정보를 확정하기 위한 방법 및 장치
CN113177980B (zh) 用于自动驾驶的目标对象速度确定方法、装置及电子设备
CN114676178A (zh) 事故检测方法、装置及电子设备
WO2024021541A1 (zh) 目标跟踪方法、装置、设备及介质
RU2461019C1 (ru) Способ координатно-связного опознавания с применением статистической оценки разности пространственных координат
CN115168527A (zh) 实时轨迹数据处理方法、装置、系统及电子设备
CN112507957B (zh) 一种车辆关联方法、装置、路侧设备及云控平台
KR20220002064A (ko) 이미지를 처리하는 방법, 장치, 기기, 저장 매체 및 프로그램
WO2024007569A1 (zh) 航位预测方法、装置、设备及介质
CN111582543A (zh) 预测模型的生成方法、估计到达时间确定方法及装置
CN113380031B (zh) 道路信息的确定方法和装置
CN114581751B (zh) 图像识别模型的训练方法和图像识别方法、装置
CN116416500B (zh) 图像识别模型训练方法、图像识别方法、装置及电子设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23844760

Country of ref document: EP

Kind code of ref document: A1