阅读说明：本技术 目标存在概率的计算方法、装置、电子设备及存储介质 (Target existence probability calculation method and device, electronic equipment and storage medium ) 是由 丁磊 杨天锡 褚世冲 于 2020-07-03 设计创作，主要内容包括：本申请实施例提供一种目标存在概率的计算方法、装置、电子设备及存储介质。具体实现方案为：接收传感器报送的目标检测数据；根据目标检测数据得到目标探测概率Pd；根据传感器视场建立传感器FOV概率模型；根据数据关联算法,选出与目标航迹最近的量测作为滤波器输入,滤波后得到新息协方差矩阵S,根据目标探测概率Pd、新息协方差矩阵S以及传感器FOV概率模型,计算目标存在概率。本申请实施例计算得到的目标存在概率能够反映传感器FOV概率P<Sub>S</Sub>和目标探测概率Pd的变化,尤其能反映有效测量值个数对目标存在概率的影响,即有效测量值个数多时比测量值少时的目标存在概率大。(The embodiment of the application provides a method and a device for calculating a target existence probability, electronic equipment and a storage medium. The specific implementation scheme is as follows: receiving target detection data reported by a sensor; obtaining target detection probability Pd according to target detection data; establishing a sensor FOV probability model according to a sensor field of view; according to a data association algorithm, selecting the measurement closest to the target track as the input of a filter, filtering to obtain an innovation covariance matrix S, and calculating the existence probability of the target according to the target detection probability Pd, the innovation covariance matrix S and a sensor FOV probability model. The target existence probability obtained through calculation in the embodiment of the application can reflect the FOV probability P of the sensor S And the change of the target detection probability Pd, especially reflecting the influence of the number of effective measured values on the existence probability of the target, i.e. the effective measured valuesWhen the number is large, the probability of existence of the target is larger than when the number is small.)

1. A method for calculating a probability of existence of an object, comprising:

receiving target detection data reported by a sensor;

obtaining a target detection probability Pd according to the target detection data;

establishing a sensor FOV probability model according to a sensor field of view;

according to a data association algorithm, selecting the measurement closest to the target track as filter input to obtain an innovation covariance matrix S, and calculating the existence probability of the target according to the target detection probability Pd, the innovation covariance matrix S and a sensor FOV probability model.

2. The method of claim 1, wherein obtaining a target detection probability Pd based on the target detection data comprises:

and screening the target detection data by using the Mahalanobis distance d to obtain effective measurement.

3. The method of claim 2, wherein obtaining the target detection probability Pd from the target detection data comprises:

acquiring detection probability corresponding to the effective measurement;

and taking the maximum value in the detection probability corresponding to the effective measurement as a target detection probability Pd.

4. The method of claim 3, wherein obtaining the probing probability corresponding to the valid measurement comprises:

acquiring a detection probability corresponding to an effective measurement from the target detection data under the condition that the target detection data comprises a detection probability corresponding to each measurement;

and under the condition that the target detection data does not comprise the detection probability corresponding to each measurement, taking a preset default value as the detection probability corresponding to the effective measurement.

5. The method according to any one of claims 1 to 4, wherein the data correlation algorithm method comprises an LNN data correlation algorithm, JPDA or Cheap-JPDA data correlation algorithm.

6. The method according to any of claims 1 to 4, characterized in that the filter comprises a classical Kalman filter KF, an extended Kalman filter EKF, an unscented Kalman filter UKF or an interactive multi-model tracker IMM.

7. The method according to any one of claims 1 to 4, further comprising:

the probability of the object in the sensor field of view is determined based on the position of the object in the sensor field of view and the sensor FOV probability model.

8. The method of any one of claims 1 to 4, wherein calculating an object presence probability based on the object detection probability Pd, the innovation covariance matrix S, and a sensor FOV probability model comprises calculating an object presence probability using the following formula:

d²＝v^TS^-1v

wherein LRm represents a measured likelihood ratio; pd represents the target detection probability; lambda is a clutter parameter;a likelihood function representing an ith measurement within the target wave gate; k represents the number of frames; m represents the total number of measurements; s represents an innovation covariance matrix in target tracking Kalman filtering; det (S) denotes a determinant of S; d represents the mahalanobis distance; v denotes an innovation vector, v^TRepresents a transposition of v; LR represents a target likelihood ratio; ps represents the probability of a target in the sensor field of view; pc is used to control parameters of the target PoE descent process when there is no measurement; PoE denotes the target presence probability.

9. An apparatus for calculating a probability of existence of an object, comprising:

the receiving module is used for receiving target detection data reported by the sensor;

the first processing module is used for obtaining a target detection probability Pd according to the target detection data;

the modeling module is used for establishing a sensor FOV probability model according to the sensor field of view;

and the target track management module is used for selecting the measurement closest to the target track as the input of a filter according to a data association algorithm to obtain an innovation covariance matrix S, and calculating the existence probability of the target according to the target detection probability Pd, the innovation covariance matrix S and a sensor FOV probability model.

10. The apparatus of claim 9, wherein the first processing module is configured to:

and screening the target detection data by using the Mahalanobis distance d to obtain effective measurement.

11. The apparatus of claim 10, wherein the first processing module comprises:

an obtaining submodule, configured to obtain a detection probability corresponding to the effective measurement;

and the processing submodule is used for taking the maximum value in the detection probability corresponding to the effective measurement as the target detection probability Pd.

12. The apparatus of claim 11, wherein the acquisition sub-module is configured to:

acquiring a detection probability corresponding to an effective measurement from the target detection data under the condition that the target detection data comprises a detection probability corresponding to each measurement;

and under the condition that the target detection data does not comprise the detection probability corresponding to each measurement, taking a preset default value as the detection probability corresponding to the effective measurement.

13. The apparatus of any of claims 9 to 12, wherein the data correlation algorithm method comprises an LNN data correlation algorithm, a JPDA or a Cheap-JPDA data correlation algorithm.

14. The apparatus according to any of the claims 9 to 12, characterized in that the filter comprises a classical kalman filter KF, an extended kalman filter EKF, an unscented kalman filter UKF or an interactive multi-model tracker IMM.

15. The apparatus according to any of claims 9 to 12, further comprising a second processing module for:

the probability of the object in the sensor field of view is determined based on the position of the object in the sensor field of view and the sensor FOV probability model.

16. The apparatus of any one of claims 9 to 12, wherein the target track management module is configured to calculate a target existence probability using the following formula:

d²＝v^TS^-1v

wherein LRm represents a measured likelihood ratio; pd represents the target detection probability; lambda is a clutter parameter;

17. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 8.

18. A computer readable storage medium having stored therein computer instructions which, when executed by a processor, implement the method of any one of claims 1 to 8.

Technical Field

The present application relates to the field of moving target detection technologies, and in particular, to a method and an apparatus for calculating a target existence probability, an electronic device, and a storage medium.

Background

The target existence probability is an important attribute of target tracking and represents how likely the tracked target really exists. When the target is judged to be a false target, the target Existence Probability (PoE) is an important reference index. The existing target existence probability calculation method has the following defects:

1) the values of the PoE reflect the changes of the target probability (Ps) and the target detection probability (Pd) in the sensor field of view insufficiently;

2) PoE resolution is poor.

Disclosure of Invention

The embodiment of the application provides a method and a device for calculating a target existence probability, an electronic device and a storage medium, which are used for solving the problems in the related art, and the technical scheme is as follows:

in a first aspect, an embodiment of the present application provides a method for calculating a target existence probability, where the method includes:

receiving target detection data reported by a sensor;

obtaining target detection probability Pd according to target detection data;

establishing a sensor FOV probability model according to a sensor field of view;

according to a data association algorithm, selecting the measurement closest to the target track as filter input to obtain an innovation covariance matrix S, and calculating the existence probability of the target according to the target detection probability Pd, the innovation covariance matrix S and a sensor FOV probability model.

In a second aspect, an embodiment of the present application provides an apparatus for calculating a target existence probability, including:

the receiving module is used for receiving target detection data reported by the sensor;

the first processing module is used for obtaining a target detection probability Pd according to target detection data;

the modeling module is used for establishing a sensor FOV probability model according to the sensor field of view;

and the target track management module is used for selecting the measurement closest to the target track as the input of a filter according to a data association algorithm to obtain an innovation covariance matrix S, and calculating the existence probability of the target according to the target detection probability Pd, the innovation covariance matrix S and the sensor FOV probability model.

In a third aspect, an embodiment of the present application provides an electronic device, including:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to cause the at least one processor to perform a method provided by any one of the embodiments of the present application.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium storing computer instructions that, when executed on a computer, perform a method in any one of the above-described aspects.

The advantages or beneficial effects in the above technical solution at least include: the calculated target existence probability can reflect the FOV probability P of the sensor_SAnd the change of the target detection probability Pd, and particularly reflects the influence of the number of effective measured values on the target existence probability, namely, the target existence probability is higher when the number of effective measured values is more than that when the number of the effective measured values is less than that of the effective measured values.

The foregoing summary is provided for the purpose of description only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the present application will be readily apparent by reference to the drawings and following detailed description.

Drawings

In the drawings, like reference numerals refer to the same or similar parts or elements throughout the several views unless otherwise specified. The figures are not necessarily to scale. It is appreciated that these drawings depict only some embodiments in accordance with the disclosure and are therefore not to be considered limiting of its scope.

FIG. 1 is a flow chart of a method of calculating a target presence probability according to an embodiment of the present application;

fig. 2 is a block diagram of PoE calculation in a multi-sensor fusion scenario according to another embodiment of the present application;

FIG. 3 is a schematic diagram of a sensor FOV probability model of a method for calculating a probability of presence of an object according to another embodiment of the present application;

FIG. 4 is a schematic view of an elliptic wave gate of a method for calculating a target presence probability according to another embodiment of the present application;

FIG. 5 is a diagram illustrating simulation results of an IPDA-based target existence probability calculation method according to the prior art;

FIG. 6 is a schematic diagram of a device for calculating a target presence probability according to an embodiment of the present application;

FIG. 7 is a diagram of a first processing module of a device for calculating a target presence probability according to another embodiment of the present application;

FIG. 8 is a schematic diagram of a device for calculating a target presence probability according to another embodiment of the present application;

fig. 9 is a block diagram of an electronic device for implementing a method for calculating a target presence probability according to an embodiment of the present application.

Detailed Description

In the following, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present application. Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive.

Fig. 1 is a flowchart of a method for calculating a target existence probability according to an embodiment of the present application. As shown in fig. 1, the method for calculating the target existence probability may include:

step S110, receiving target detection data reported by a sensor;

step S120, obtaining a target detection probability Pd according to target detection data;

step S125, establishing a sensor FOV (Field of View) probability model according to the sensor Field of View;

step S130, selecting the measurement (measures) closest to the target track as filter input according to a data association algorithm to obtain an innovation covariance matrix S, and calculating the target existence probability PoE according to the target detection probability Pd, the innovation covariance matrix and the sensor FOV probability model.

Fig. 2 is a block diagram of PoE calculation in a multi-sensor fusion scenario according to another embodiment of the present application. In the example of fig. 2, the sensor includes a millimeter wave radar and a camera. Fig. 2 shows a calculation flow and a calculation module of the existence probability of the target after the millimeter wave radar and the camera are fused. The calculation module comprises a data association module 1, an IMM (Interactive Multiple Model algorithm) tracker 2, a sensor FOV probability Model 3 and a PoE calculation module. The data association module, the IMM tracker and the sensor FOV probability model are input modules of PoE calculation, and the PoE calculation module is an output module.

Referring to fig. 1 and 2, target detection data reported from the sensor is received in step S110. In the example of fig. 2, the sensor includes a millimeter wave radar and a camera. M1 and M2 in fig. 2 represent target measurement values reported by the millimeter wave radar and the camera, respectively.

In step S120, the target detection probability Pd is obtained according to the target detection data reported by the sensor. In the application scenario of single-sensor target tracking, Pd depends on the signal-to-noise ratio SNR, power, false alarm rate (Pfa), etc. of the sensor itself. In the application scenario of multi-sensor fusion, the measurements (target detection data) from at least two sensors are associated with the established target track, and the target detection probability Pd is obtained from the measurements within the gate. The data association module 1 in the example of fig. 2 is configured to associate the millimeter wave radar and the camera reported magnitude, and determine the target detection probability Pd by using a predetermined algorithm. The Pd is then transmitted to the PoE computation module.

In step S130, the measurement with the highest correlation with the target track is selected as the filter input from the target detection data. In one example, the degree of association may specifically be represented by a mahalanobis distance d. The filter may employ an IMM tracker. Referring to fig. 2, in the data association module, a Local Nearest Neighbor algorithm (LNN) may be used to select a measurement with the maximum association degree with the target track from the target detection data, and transmit the relevant measured data to the IMM tracker to update the input data of the IMM tracker. And then obtaining an innovation covariance matrix S corresponding to the measurement by using an IMM tracker. And finally, transmitting the innovation covariance matrix S to a PoE calculation module.

In one example, the data correlation module may use a local nearest neighbor algorithm and use an elliptic wave gate to select the metric from the target detection data that has the greatest correlation to the target track. The determination criterion of the wave gate is that the Mahalanobis distance of each measurement obeys χ if the measurements are independent²And (4) distribution. Let the probability of the real measurement falling in the track wave gate be P_GUsually taking P_GNot less than 0.95, and the measurement dimension is χ²The degree of freedom of distribution is according to P_GAnd measure dimension chi²The table may determine a gate threshold for the mahalanobis distance. For example, let P_GIs 0.99, m is 2, find chi²Table, when the degree of freedom is 2 and the error probability is 1-P_GWhen the value is equal to 0.01, the value of gamma is equal to 9.21

d²<γ＝9.21

That is, when the mahalanobis distance d < sqrt (γ) ═ 3.03 for each measurement and track, the measurement falling within the wave gate is considered to be an effective measurement.

Besides the elliptic wave gate, a rectangular wave gate, an annular wave gate, a sector wave gate in a polar coordinate system, and the like can also be applied in the embodiment of the application.

Referring to the example of FIG. 2, in one embodiment, the tracker includes a Kalman filter based on an interactive multimodal algorithm IMM.

In one example, the IMM tracker contains two models, CV (Constant Velocity), CA (Constant Acceleration), and Kalman filter.

In step S125, a sensor FOV probability model is established based on the sensor field of view. In step S130, a target existence probability (PoE) is calculated based on the target detection probability Pd obtained in step S120, the innovation covariance matrix S, and the target probability Ps in the sensor field. Wherein, the object probability Ps in the sensor field of view is determined according to the position of the object in the sensor field of view and the sensor FOV probability model.

Referring to the example of fig. 2, a sensor FOV probability model may be used to derive a target probability Ps in the sensor field of view and transmit Ps to the PoE calculation module. And finally, calculating by a PoE calculation module according to the target detection probability Pd, a target probability model Ps of the FOV of the sensor field of view and the innovation covariance matrix to obtain the target existence probability PoE.

The target existence probability obtained through calculation in the embodiment of the application can reflect the FOV probability P of the sensor_SAnd the change of the target detection probability Pd, and particularly reflects the influence of the number of effective measured values on the target existence probability, namely, the target existence probability is higher when the number of effective measured values is more than that when the number of the effective measured values is less than that of the effective measured values.

In one embodiment, the method further comprises: and determining the probability Ps of the target in the sensor field of view according to the position of the target in the sensor field of view and the sensor FOV probability model.

Fig. 3 is a schematic diagram of a sensor FOV probability model of a method for calculating a target presence probability according to another embodiment of the present application. Referring to fig. 3, still taking radar and camera as an example, as a prior probability, an exemplary partitioning rule of the existence probability Ps of the target in the FOV field is as follows:

1) the Ps corresponding to the central FOV area (boresight) of the radar or camera is large, and the Ps corresponding to the peripheral FOV area is small. That is, if the position of the target in the sensor field of view is located in the central FOV area, the value of Ps in this case is large. In fig. 3, a sector area within 10 ° indicates the detection range of the radar, and a sector area within 50 ° indicates the detection range of the camera. As can be seen from fig. 3, the smaller the angle, the closer to the central FOV area, the larger the corresponding value of Ps.

2) Ps corresponding to the area covered by the radar and the camera is larger than Ps corresponding to the single sensor coverage area. As can be seen from fig. 3, the coverage area of the radar and the camera jointly corresponds to Ps of 0.9, the coverage area of the radar single sensor corresponds to Ps of 0.8 and 0.85, and the coverage area of the camera single sensor corresponds to Ps of 0.7 and 0.75. It can be seen that the common coverage area corresponds to a greater Ps than the single sensor coverage area.

3) Ps corresponding to the near region of the radar or camera is larger than Ps corresponding to the far region. The reason for this is that the SNR (Signal to Noise Ratio) is generally higher in the near than in the far, and the reflected power is higher in the near than in the far for the radar. As can be seen from fig. 3, in the case of the same detection angle, Ps corresponding to the near region is larger than Ps corresponding to the far region. Taking radar as an example, Ps corresponding to the near area is 0.9, and Ps corresponding to the far area is 0.8 and 0.85, respectively.

In one example, a Local Nearest Neighbor algorithm (LNN) may be used to select the metric from the target detection data that has the greatest correlation to the target track. For example, the measurement closest to the mahalanobis distance of the target track is selected from the target detection data, and then the other measurement data is discarded, and the measurement is used as the target, and the position of the target in the sensor field of view and a preset division rule are used to determine the target probability Ps in the sensor field of view.

In one embodiment, obtaining the target detection probability Pd according to the target detection data includes:

and screening the target detection data by using the Mahalanobis distance d to obtain effective measurement.

In the embodiment of the application, for each target track, on one hand, the measurement transmitted from the millimeter wave radar and the camera is associated with the target track, for example, a local nearest neighbor algorithm LNN is used as a data association method; on the other hand, the detection probability of each measurement in the elliptic wave gate determined by the mahalanobis distance threshold can be calculated. And determining that the measurement is an effective measurement falling within the wave gate under the condition that the mahalanobis distance between each measurement and the track is smaller than a preset mahalanobis distance threshold value.

Referring to fig. 2, the data correlation module receives the measurement values M1, M2, …, Mn from the sensors, and uses mahalanobis distance as the fine wave gate for screening, and uses the measurement in the wave gate as the effective measurement.

In one embodiment, obtaining the target detection probability Pd according to the target detection data includes:

acquiring detection probability corresponding to effective measurement;

and taking the maximum value in the detection probability corresponding to the effective measurement as the target detection probability Pd.

In one embodiment, obtaining the detection probability corresponding to the effective measurement includes:

acquiring a detection probability corresponding to an effective measurement from target detection data under the condition that the target detection data comprises a detection probability corresponding to each measurement;

and under the condition that the target detection data does not comprise the detection probability corresponding to each measurement, taking a preset default value as the detection probability corresponding to the effective measurement.

The probability of detection for each measurement is determined by SNR, power, false alarm rate (Pfa), etc. If Pfa is the same, the higher the SNR and the higher the power, the higher the detection probability corresponding to the measurement. Fig. 4 is a schematic view of an elliptic wave gate of a method for calculating a target existence probability according to another embodiment of the present application. As shown in fig. 4, still taking radar and camera as examples, an exemplary method for obtaining a target detection probability according to the target detection data may include:

1) if a value of the detection probability corresponding to each measurement is provided in the target detection data output by the radar or the camera, the target detection probability Pd is determined using the value. If the value of the detection probability corresponding to each measurement is not provided in the target detection data output by the radar or the camera, the default value is used as the value of the detection probability corresponding to each measurement, and the target detection probability Pd is determined by using the value. For example, a default value of 0.85 may be set.

2) If there are multiple targets output by the sensors in the gate, each target may correspond to one measurement or multiple measurements, and multiple targets correspond to multiple measurements, then the maximum value from the values of the detection probabilities corresponding to the multiple measurements is taken as the target detection probability Pd. The formula for calculating Pd can be expressed as:

Pd＝max(Pd1，Pd2，…，Pdm)

wherein n is the number measured in the wave gate, and Pd1, Pd2, … and Pdm respectively represent the detection probability corresponding to each measurement in the wave gate.

In the embodiment of the present application, when one measured detection probability represents the target detection probability Pd of each measurement in the wave gate, in addition to the method of taking the maximum value in the above example, an average value may be used, that is, an average value of the detection probabilities corresponding to a plurality of measurements is taken as the target detection probability Pd.

In the above example, there are multiple targets in the gate that the sensor outputs. Referring to fig. 2 again, the data association module 1 is configured to associate target detection data reported by a plurality of sensors, and determine a target detection probability Pd according to data corresponding to a plurality of measurements of the target detection data by using a predetermined algorithm. The predetermined algorithm may include the above calculation formula for Pd.

Alternatively, if the gate has only one target output from the sensor, and one target may correspond to one measurement or a plurality of measurements, the maximum value is taken as the target detection probability Pd from the values of the detection probability corresponding to the measurement or the plurality of measurements. Embodiments of the application may be used to calculate the confidence or confidence of a single sensor target. Such as for calculating the confidence of a target during millimeter wave radar target tracking or camera image target tracking. At the moment, Ps is also related to a FOV probability model of the sensor, and the center of the FOV is large in Ps, the edge of the FOV is small in Ps, the near point of the FOV is large in Ps, and the far point of the FOV is small in Ps. Pd is related to SNR, distance, etc. of the target, and tracking performance is related to mahalanobis distance.

In one embodiment, the Data Association algorithm method includes an LNN Data Association algorithm, a JPDA or a chemically-JPDA (simple-Joint Probabilistic Data Association) Data Association algorithm. The basic idea of associating JPDA with joint probability data is as follows: all measurements within the drop-in gate are associated with different probabilities with the track, i.e. all measurements participate in the track update with different weights (probabilities). The Cheap-JPDA is a JPDA modifying algorithm, and overcomes the defect of large computation amount of the JPDA at the cost of sacrificing a little accuracy of data association.

In one embodiment, the filter comprises a classical kalman filter KF, an extended kalman filter EKF, an unscented kalman filter UKF, or an interactive multi-model tracker IMM.

In one embodiment, calculating the target existence probability according to the target detection probability Pd, the innovation covariance matrix S, and the sensor FOV probability model includes calculating the target existence probability using the following formula:

d²＝v^TS^-1v

wherein LRm denotes a measurement likelihood ratio (measurement likelihood ratio); pd represents the target detection probability; lambda is a clutter parameter, and the bigger the lambda value is, the denser the clutter is;a measurement likelihood (measurement likelihood) is represented, and specifically a likelihood function, namely a probability density, of the ith measurement (target reported by the sensor) in the target wave gate is represented; k represents the number of frames; m represents the total number of measurements; s represents an innovation covariance matrix in target tracking Kalman filtering; det (S) denotes a determinant of S; d represents the Mahalanobis distance (Mahalanobis distance); v denotes an innovation vector, v^TRepresents a transposition of v; LR represents a target likelihood ratio (target likelihood ratio); ps represents the probability of a target in the sensor field of view; pc is used to control parameters of the target PoE descent process when there is no measurement; PoE denotes target presence probability, with an initial value set to 0.5.

The PoE calculation method is also referred to as IPDA (Integrated Probabilistic Data Association) method. The LR calculation formula includes "presence measurement" and "no measurement". "no measurement" includes the case where the detection data is not obtained due to equipment failure or the like; the "presence measurement" includes a case where the detection data can be normally acquired.

PoE indexes designed in application scenarios such as fusion by using a plurality of sensors or target tracking of a single sensor can generally meet the following design requirements:

r1) the FOV (Field of View) of the sensor is larger in the central area than in the edge area.

R2), if the target detection probability (expressed by Pd) is higher, the corresponding PoE is higher, that is, the PoE has good resolution to Pd.

R3) for the same wave gate, the PoE after integration of multiple measurements (measures) is greater than the PoE of a single measurement.

Wherein for multi-sensor fusion applications, metrology is the target of sensor reporting; for a single sensor target tracking application scenario, metrology refers to the measurement data entering the filter. The actual target may produce one or more measurements.

R4) during the object establishment, PoE should be increased step by step; in the process of target disappearance, PoE is gradually reduced; in the target tracking process, the tracking performance is better, namely the smaller the mahalanobis distance is, the larger the PoE is.

Wherein the sensor may comprise a radar or a camera. Mahalanobis distance represents the statistical distance between the measurement and the track.

The PoE calculation method can well meet the PoE design requirements R1-R4.

The target existence probability results obtained by calculation in the embodiment of the application can be modeled and simulated in MATLAB. In one example, targets for 5 different Pd may be set, using an IMM kalman tracker, for data correlation using LNN. The simulated total frame data is 1200 frames. In the previous 400 frames, the probability of objects Ps in the sensor field of view is gradually increased, and 5 objects with different Pd values cut into the sensor FOV. In the middle 300 frames, the value of the target probability Ps in the sensor field of view is the largest. In the last 500 frames, the target probability Ps in the sensor field of view gradually decreases. The target interruption is simulated at 300-. The simulation result of the target existence probability calculated in the embodiment of the present application is shown in fig. 5.

The upper left diagram and the lower left diagram in fig. 5 represent simulation conditions. In the upper left diagram in fig. 5, the abscissa t [ k ] represents a frame; the ordinate dis _ Ma represents the Mahalanobis Distance of the sliding window (Mahalanobis Distance with sliding window); the dark images represent the corresponding images without filtering (raw dis _ Ma); the light color image represents an image corresponding to the case after filtering (filtered dis _ Ma). In the lower left diagram in fig. 5, the abscissa t [ k ] represents a frame; the ordinate represents Ps or Ps × Pd determined from a Sensor model.

In the upper right diagram in fig. 5, the abscissa t [ k ] represents a frame; the ordinate LR _ IPDA represents the value of LR in the target existence probability calculation method of IPDA. The image on the upper right in fig. 5 represents an image (LRwith differential Pd) in which the value of LR changes with a Different Pd value.

In the lower right diagram in fig. 5, the abscissa t [ k ] represents a frame; the ordinate indicates the value of POE. The image in the lower right of fig. 5 shows an image in which the value of PoE obtained by LR changes with a different Pd value (PoE of LR with differential Pd).

Referring to the simulation result shown in fig. 5, the method for calculating the target existence probability according to the embodiment of the present application has the following advantages:

1) PoE can reflect changes in Ps and Pd.

2) The expression of the target likelihood ratio LR and the formula for converting the expression into the target existence probability are designed, so that different PoE can be obtained after the targets corresponding to different Pd and Ps are calculated. In LRm expressionAnd summing and accumulating calculation is carried out, so that the measured number of PoE is obviously reflected, and the PoE with more effective measured number is larger than the PoE with less effective measured number.

FIG. 6 is a diagram illustrating an apparatus for calculating a target presence probability according to an embodiment of the present application. As shown in fig. 6, the calculating means of the target existence probability may include:

the receiving module 100 is used for receiving target detection data reported by the sensor;

the first processing module 200 is configured to obtain a target detection probability Pd according to target detection data;

the modeling module 220 is used for establishing a sensor FOV probability model according to the sensor field of view;

and the target track management module 400 is configured to select a measurement closest to the target track as a filter input according to a data association algorithm to obtain an innovation covariance matrix S, and calculate a target existence probability according to the target detection probability Pd, the innovation covariance matrix S, and the sensor FOV probability model.

In one embodiment, the first processing module 200 is configured to:

and screening the target detection data by using the Mahalanobis distance d to obtain effective measurement.

Fig. 7 is a schematic diagram of a first processing module of a device for calculating a target existence probability according to another embodiment of the present application. As shown in fig. 7, in one embodiment, the first processing module 200 includes:

an obtaining sub-module 210, configured to obtain a detection probability corresponding to the effective measurement;

and the processing sub-module 220 is configured to use a maximum value of the detection probabilities corresponding to the effective measurement as the target detection probability Pd.

In one embodiment, the obtaining sub-module 210 is configured to:

acquiring a detection probability corresponding to an effective measurement from target detection data under the condition that the target detection data comprises a detection probability corresponding to each measurement;

and under the condition that the target detection data does not comprise the detection probability corresponding to each measurement, taking a preset default value as the detection probability corresponding to the effective measurement.

In one embodiment, the data correlation algorithm method comprises an LNN data correlation, JPDA or Cheap-JPDA data correlation algorithm.

In one embodiment, the filter comprises a classical kalman filter KF, an extended kalman filter EKF, an unscented kalman filter UKF, or an interactive multi-model tracker IMM.

Fig. 8 is a schematic diagram of a device for calculating a target existence probability according to another embodiment of the present application. As shown in fig. 8, in one embodiment, the apparatus further includes a second processing module 350 configured to:

the probability of the object in the sensor field of view is determined based on the position of the object in the sensor field of view and the sensor FOV probability model.

In one embodiment, the target track management module 400 is configured to calculate the target existence probability using the following formula:

d²＝v^TS^-1v

wherein LRm denotes a measurement likelihood ratio (measurement likelihood ratio); pd represents the target detection probability; lambda is a clutter parameter, and the bigger the lambda value is, the denser the clutter is;a measurement likelihood (measurement likelihood) is represented, and specifically a likelihood function, namely a probability density, of the ith measurement (target reported by the sensor) in the target wave gate is represented; k represents the number of frames; m represents the total number of measurements; s represents an innovation covariance matrix in target tracking Kalman filtering; det (S) denotes a determinant of S; d represents the Mahalanobis distance (Mahalanobis distance); v denotes an innovation vector, v^TRepresents a transposition of v; LR represents a target likelihood ratio (target likelihood ratio); ps represents the probability of a target in the sensor field of view; pc is used to control parameters of the target PoE descent process when there is no measurement; PoE denotes the target presence probability.

The functions of each module in the target existence probability calculation apparatus according to the embodiment of the present application may refer to the corresponding descriptions in the above method, and are not described herein again.

Fig. 9 is a block diagram of an electronic device for implementing a method for calculating a target presence probability according to an embodiment of the present application. As shown in fig. 9, the control apparatus includes: a memory 910 and a processor 920, the memory 910 having stored therein instructions executable on the processor 920. The processor 920, when executing the instructions, implements the method of calculating the target presence probability in the above embodiments. The number of the memory 910 and the processor 920 may be one or more. The control device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The control device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present application that are described and/or claimed herein.

The control device may further include a communication interface 930 for communicating with an external device for data interactive transmission. The various devices are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor 920 may process instructions for execution within the control device, including instructions stored in or on a memory to display graphical information of a GUI on an external input/output apparatus (such as a display device coupled to an interface). In other embodiments, multiple processors and/or multiple buses may be used, along with multiple memories and multiple memories, as desired. Also, multiple control devices may be connected, with each device providing portions of the necessary operations (e.g., as a server array, a group of blade servers, or a multi-processor system). The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown in FIG. 9, but this does not indicate only one bus or one type of bus.

Optionally, in an implementation, if the memory 910, the processor 920 and the communication interface 930 are integrated on a chip, the memory 910, the processor 920 and the communication interface 930 may complete communication with each other through an internal interface.

It should be understood that the processor may be a Central Processing Unit (CPU), other general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. A general purpose processor may be a microprocessor or any conventional processor or the like. It is noted that the processor may be an advanced reduced instruction set machine (ARM) architecture supported processor.

Embodiments of the present application provide a computer-readable storage medium (such as the above-mentioned memory 910) storing computer instructions, which when executed by a processor implement the methods provided in embodiments of the present application.

Alternatively, the memory 910 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the calculation means of the target existence probability, and the like. Further, the memory 910 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 910 may optionally include memory located remotely from the processor 920, and such remote memory may be connected to the target probability presence computing device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.

Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present application, "a plurality" means two or more unless specifically limited otherwise.

Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more (two or more) executable instructions for implementing specific logical functions or steps in the process. And the scope of the preferred embodiments of the present application includes other implementations in which functions may be performed out of the order shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved.

The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions.

It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. All or part of the steps of the method of the above embodiments may be implemented by hardware that is configured to be instructed to perform the relevant steps by a program, which may be stored in a computer-readable storage medium, and which, when executed, includes one or a combination of the steps of the method embodiments.

In addition, functional modules in the embodiments of the present application may be integrated into one processing module, or each module may exist alone physically, or two or more modules are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module may also be stored in a computer-readable storage medium if it is implemented in the form of a software functional module and sold or used as a separate product. The storage medium may be a read-only memory, a magnetic or optical disk, or the like.

While the present invention has been described with reference to the preferred embodiments, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.