A target detection method and system

Abstract

The target detection method and system disclosed by the application directly utilize the corresponding state prediction value to calculate the difference result after obtaining the state measurement value of the target at the current time, judge whether the state measurement value is a singular value according to the calculation result, correct the measurement value in the case of judging that it is a singular value, and determine the state correction value of the target according to the correction result. Therefore, the application can identify whether the measurement value is a singular value in real time, and can correct the measurement value judged as a singular value in real time, thereby greatly reducing the probability of the occurrence of the inaccurate detection result due to the existence of the singular value, and improving the track continuity and stability in the target detection process.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a target detection method and system. Background Technology

[0002] Intelligent technologies in automobiles are a crucial means of addressing traffic safety issues. Target detection and tracking are indispensable components of intelligent driving assistance systems (ADAS), and the accuracy, consistency, and real-time performance of target detection and tracking in these systems significantly impact their overall performance. Millimeter-wave radar is typically used for target detection in ADAS. Its superior speed measurement capabilities and excellent penetration through rain and fog make it an irreplaceable sensor choice for ADAS solutions.

[0003] However, due to the limitations of millimeter-wave radar's resolution and the size of the target it detects, outliers inevitably occur during target detection. These outliers can cause sudden jumps in the target's trajectory, which greatly affects the continuity and stability of the trajectory during target detection. Summary of the Invention

[0004] In view of this, the present invention provides a target detection method and system to solve the problem that singular values ​​affect the continuity and stability of the track during the target detection process in the prior art.

[0005] A target detection method, comprising:

[0006] Obtain the target's state measurement value at the current moment;

[0007] Determine the predicted state value corresponding to the measured state value;

[0008] Calculate the difference between the predicted state value and the measured state value;

[0009] Based on the difference results and the preset discrimination rules, it is determined whether the state measurement value is an outlier.

[0010] If the state measurement value is an outlier, the state measurement value is corrected.

[0011] The state correction value of the target is obtained based on the corrected state measurement value.

[0012] Optionally, the calculation of the difference between the predicted state value and the measured state value includes:

[0013] The measured value of the physical quantity to be compared is determined using the state measurement value; the physical quantity to be compared is used to describe the state of the target.

[0014] The predicted value of the physical quantity to be compared is determined using the state prediction value;

[0015] Calculate the difference between the measured value and the predicted value of the physical quantity to be compared, and use the difference value as the difference result.

[0016] Optionally, when the coordinate system of the predicted state value is a rectangular coordinate system and the coordinate system of the measured state value is a polar coordinate system, the physical quantities to be compared include: Doppler velocity, longitudinal displacement, and lateral displacement.

[0017] The step of calculating the difference between the measured value and the predicted value of the physical quantity to be compared, and using the difference as the difference result, includes:

[0018] The magnitudes of the differences between the measured and predicted lateral displacement values, the magnitudes of the differences between the measured and predicted longitudinal displacement values, and the magnitudes of the differences between the predicted and measured Doppler velocity values ​​are used as the difference results between the measured and predicted state values.

[0019] Optionally, the process of determining whether the state measurement value is a singular value based on the difference result and the preset discrimination rule includes:

[0020] If the difference result satisfies a preset discrimination rule, the state measurement value is identified as an outlier. The preset discrimination rule includes: the difference result exceeds a preset range, where the preset range is an allowable range representing the difference between the state measurement value and the state prediction value.

[0021] Optionally, the difference results satisfy a preset discrimination rule, and identifying the state measurement value as an outlier includes:

[0022] If at least one of the following exceeds its corresponding preset range: the magnitude of the difference between the lateral displacement measurement value and the lateral displacement prediction value, the magnitude of the difference between the longitudinal displacement measurement value and the longitudinal displacement prediction value, and the magnitude of the difference between the Doppler velocity prediction value and the Doppler velocity measurement value, then the state measurement value is identified as an outlier. The physical quantity to be compared whose difference magnitude exceeds the corresponding preset range is recorded as an outlier physical quantity to be compared, and the value of the outlier value discrimination counter corresponding to the outlier physical quantity to be compared is incremented by 1. The value is used to characterize the number of consecutive cumulative times of the outlier physical quantity to be compared.

[0023] Optionally, correcting the state measurement value includes:

[0024] The first correction coefficient of the corresponding singular physical quantity to be compared is determined based on the value of each singular value discrimination counter;

[0025] The measured values ​​of the singular physical quantities to be compared are corrected based on the first correction factor;

[0026] The correction value of the state measurement value is determined based on the correction value of the singular physical quantity to be compared.

[0027] Optionally, determining the first correction coefficient for the corresponding singular physical quantity to be compared based on the values ​​of each singular value discrimination counter includes:

[0028] If the value of the singular value discrimination counter is not greater than a preset value, the first correction coefficient is set according to the first preset rule. The first preset rule is to reduce the modulus of the difference between the singular physical quantities to be compared after correction. Each singular physical quantity to be compared corresponds to a singular value discrimination counter.

[0029] If the value of the singular value discrimination counter is greater than a preset value, the first correction coefficient is set according to the second preset rule, wherein the second preset rule is that the modulus of the difference between the singular physical quantities to be compared remains unchanged or increases after correction.

[0030] Optionally, correcting the state measurement value includes:

[0031] Based on the value of the singular value discrimination counter, a first correction coefficient and a second correction coefficient are determined for the corresponding singular physical quantity to be compared, wherein the sum of the first correction coefficient and the second correction coefficient is 1;

[0032] The measured value of the singular physical quantity to be compared is corrected based on the first correction coefficient, and the predicted value of the physical quantity to be compared is corrected based on the second correction coefficient;

[0033] The corrected predicted value of the singular physical quantity to be compared and the sum of the corrected predicted values ​​are taken as the corrected value of the singular physical quantity to be compared.

[0034] The correction value of the state measurement value is determined based on the correction value of the singular physical quantity to be compared.

[0035] Optionally, determining the first and second correction coefficients corresponding to the physical quantity to be compared based on the value of the singularity discrimination counter includes:

[0036] If the value of the singular value discrimination counter is not greater than a preset value, then the first correction coefficient and the second correction coefficient are set according to the first preset rule, wherein the first preset rule is to reduce the modulus of the difference between the physical quantities to be compared after correction.

[0037] If the value of the singularity discrimination counter is greater than a preset value, the first correction coefficient and the second correction coefficient are set according to the second preset rule, wherein the modulus of the difference between the physical quantities to be compared remains unchanged or increases after correction.

[0038] The present invention also discloses a target detection system, comprising:

[0039] The measurement value acquisition module is used to acquire the target's state measurement value at the current moment;

[0040] The prediction value acquisition module is used to determine the state prediction value corresponding to the state measurement value;

[0041] The difference result calculation module is used to calculate the difference result between the predicted state value and the measured state value;

[0042] The singular value determination module is used to determine whether the state measurement value is a singular value based on the difference result and the preset discrimination rule;

[0043] A correction module is used to correct the state measurement value when the state measurement value is an singular value.

[0044] The state correction value acquisition module is used to obtain the state correction value of the target based on the corrected state measurement value.

[0045] In summary, the target detection method and system disclosed in this invention, after obtaining the target's state measurement value at the current moment, directly calculates the difference result using its corresponding state prediction value. Based on the calculation result, it determines whether the state measurement value is a singular value. If it is determined to be a singular value, the measurement value is corrected, and the corrected state value of the target is determined based on the correction result. Therefore, this invention can identify whether the measurement value is a singular value in real time and can correct the measurement value determined to be a singular value in real time, thereby greatly reducing the probability of inaccurate detection results due to the presence of singular values ​​and improving the continuity and stability of the track during target detection. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart of a target detection method disclosed in this invention;

[0048] Figure 2 This is a schematic diagram of the observation geometry in the polar coordinate system disclosed in this invention;

[0049] Figure 3 This is a flowchart of another target detection method disclosed in this invention;

[0050] Figure 4 This is a flowchart of another target detection method disclosed in this invention;

[0051] Figure 5 This is a schematic diagram of the target's lateral position change curve during testing, as disclosed in this invention.

[0052] Figure 6 This is a schematic diagram of the longitudinal position change curve of the target during the test disclosed in this invention;

[0053] Figure 7 This is a schematic diagram of the structure of a target detection system disclosed in this invention. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Analysis of existing target detection processes reveals that the presence of singular values ​​significantly impacts the continuity and stability of the target detection trajectory. Therefore, this invention aims to provide a target detection method, as well as an apparatus and system for implementing the method, in order to reduce the impact of singular values.

[0056] To reduce the impact of singular values, two aspects are required: first, to accurately identify singular values, that is, to discover their existence; and second, to process the discovered singular values.

[0057] Since outliers are abnormal values ​​measured during target detection, a large amount of measurement data can be collected and statistically analyzed offline. After calculating the mean and variance of a set of data, a judgment criterion is established based on the mean and variance. This criterion is used to identify outliers in the data set, and then the outliers are discarded, thus achieving outlier identification and removal from the measurement data. However, this outlier identification process cannot meet the real-time requirements of data processing, which are very demanding in intelligent driving scenarios, thus limiting its widespread application. Furthermore, directly discarding outliers is also problematic. In actual detection processes, the target's position may suddenly change during movement, and directly discarding outliers can cause discontinuity in target tracking, leading to the serious problem of target loss.

[0058] Through further research, the inventors have provided a target detection method. This method, after obtaining the target's current state measurement value, directly uses its corresponding predicted state value—that is, the pre-predicted current state value—to calculate the difference between these two values. Based on the calculation result, it directly determines whether the state measurement value is a singular value. Furthermore, if it is determined to be a singular value, the state measurement value is corrected, and the corrected state value of the target is determined based on the correction result. By executing the above process at every detection moment, it is possible to identify whether the state measurement value is a singular value in real time during target detection and to correct it in real time. This significantly reduces the probability of inaccurate detection results due to the presence of singular values, improving the continuity and stability of the trajectory during target detection.

[0059] Figure 1 The diagram shown is a schematic flowchart of a target detection method disclosed in an embodiment of the present invention, including:

[0060] Step S101: Obtain the target's state measurement value at the current moment.

[0061] In one possible implementation, target state detection is achieved through radar. In this case, the state measurement value is the polar coordinate measurement X, which is the radar's measurement of the target state at the current moment. m =[R,V d ,θ]. Where R is the radial distance, V d The value is the Doppler velocity measurement, and θ is the direction angle.

[0062] Step S102: Determine the state prediction value corresponding to the state measurement value.

[0063] This step can predict the target state X at the current moment based on the assumed target motion state model (i.e., motion state transition model) and the optimized value of the target state at the previous moment.pre Typically, the target motion state model is established in a Cartesian coordinate system, and the predicted value of the target state is X. pre =[x p ,v xp ,y p ,v yp ,] T The predicted value of the target state, also known as the state prediction value, is x. p The predicted x-coordinate of the target's location, y p v is the predicted ordinate of the target's location. xp v is the predicted lateral velocity of the target. yp The longitudinal velocity prediction value for the target.

[0064] Step S103: Calculate the difference between the predicted state value and the measured state value.

[0065] After obtaining the target's current state measurement and state prediction values, the difference result is directly calculated using these two values. This difference result forms the basis for subsequent determination of whether a value is singular.

[0066] As the analysis in the previous step shows, neither the state measurement value nor the state prediction value is a single numerical value, but rather a combination of physical quantities. In other words, the state of the target is described by these physical quantities. Therefore, the difference results actually reflect the differences between the two sets of physical quantities. The reason they can be used for difference comparison is based on the premise that the physical quantities being compared represent the same physical meaning or have the same physical properties.

[0067] Therefore, it is necessary to determine the measured value of the physical quantity to be compared based on the state measurement value, determine the predicted value of the physical quantity to be compared based on the state prediction value, and then calculate the difference between the measured value of the physical quantity to be compared and the predicted value of the physical quantity to be compared, and use the difference value as the difference result.

[0068] Step S104: Based on the difference results and preset discrimination rules, determine whether the state measurement value is an outlier.

[0069] If the difference result meets the preset discrimination rule, the state measurement value is identified as an outlier. The preset discrimination rule includes: the difference result exceeds a preset range, and the preset range is the allowable range that characterizes the difference between the state measurement value and the state prediction value.

[0070] In other words, if the difference exceeds the allowable range of the difference between the state measurement value and the state prediction value, then the state measurement value at the current moment is a singular value.

[0071] The preset range in this embodiment can be set based on empirical values. Typically, a preset maximum threshold is set, and the range of values ​​from 0 to this preset maximum threshold is used as the preset range. If the difference result is within this range, it is not a singular value; if it exceeds this preset range, that is, is greater than or equal to the preset maximum threshold, the state measurement value is considered a singular value.

[0072] Step S105: Correct the state measurement value if the state measurement value is an outlier.

[0073] At this point, the state measurement value can be corrected according to some set rules. For example, the measurement value can be increased or decreased according to a preset ratio, or a correction coefficient can be set and the state measurement value can be multiplied by the correction coefficient to obtain the corrected state measurement value, or it can also be called the optimized state measurement value.

[0074] Step S106: Obtain the target's state correction value based on the corrected state measurement value.

[0075] Algorithms such as the extended Kalman filter and the unscented Kalman filter are used to filter the state measurement values, and the results are used as the state correction value of the target at the current moment, that is, the optimized value of the target state, to complete the target detection process at the current moment.

[0076] When the next moment arrives, repeat steps S101-106. The state correction value of the target at this time is used as the key parameter for obtaining the state prediction value of the target in the next moment.

[0077] The target detection method disclosed in this embodiment can identify whether the state measurement value is a singular value in real time during the target detection process and correct it in real time, thereby greatly reducing the impact of singular values ​​on the accuracy of the detection results and improving the continuity and stability of the track during the target detection process.

[0078] In one possible implementation, the coordinate system for the predicted state values ​​is a Cartesian coordinate system, and the coordinate system for the measured state values ​​is a polar coordinate system. To calculate the differences between the physical quantities to be compared, the two coordinate systems need to be unified first. For example, if coordinate system one is a Cartesian coordinate system, the physical quantities to be compared can include: velocity, longitudinal displacement, and lateral displacement. Alternatively, if coordinate system one is a polar coordinate system, the physical quantities to be compared include: the Doppler velocity measurement value V. d And radial distance R. For example... Figure 2 The diagram showing the observation geometry in the polar coordinate system illustrates that there is a certain correspondence between the physical quantities to be compared between the polar coordinate system and the rectangular coordinate system, enabling coordinate transformation.

[0079] Further research by the inventors revealed that when transforming the coordinates of physical quantities, it is not necessarily necessary to completely unify them to a specific coordinate system. As long as the coordinate systems of the physical quantities being compared are the same, the difference between the two can be obtained. Moreover, the ease of calculating the difference varies for different physical quantities in different coordinate systems. To reduce the complexity of the solution implementation and improve the speed of the detection process, the inventors developed the following optional processing method:

[0080] The physical quantities characterizing the position and state of the target are unified in a rectangular coordinate system, while the physical quantities characterizing the velocity of the target are unified in polar coordinates.

[0081] The specific process includes, according to the formula, transforming the radial distance of the state measurement value in the polar coordinate system into the abscissa measurement value of the target position in the rectangular coordinate system, i.e., x. m =R·sinθ, the measured value of the ordinate y m = R·cosθ, where θ is the azimuth angle. The velocity component v in the predicted value in the rectangular coordinate system... xp and v yp Doppler velocity prediction V in polar coordinates p =v xp ·sinθ p +v yp ·cosθ p .in,

[0082] Therefore, the physical quantities to be compared include: Doppler velocity, longitudinal displacement, and lateral displacement. That is, the Doppler velocity in the polar coordinate system represents the target's velocity at the current moment, while the longitudinal and lateral displacements in the Cartesian coordinate system represent the target's position at the current moment. In the Cartesian coordinate system, the longitudinal displacement is the ordinate, and the lateral displacement is the abscissa.

[0083] Accordingly, the difference between the measured value and the predicted value of the physical quantity to be compared is calculated, and the difference is used as the difference result, including:

[0084] The magnitudes of the differences between the measured and predicted lateral displacement values, the longitudinal displacement values, and the Doppler velocity predictions are used as the difference results between the measured and predicted state values. This yields three difference results. These three results are then used to determine whether a preset discrimination rule is met. The specific methods for identifying singular values ​​in the state measurements include:

[0085] If at least one of the following exceeds its corresponding preset range: the magnitude of the difference between the lateral displacement measurement value and the lateral displacement prediction value, the magnitude of the difference between the longitudinal displacement measurement value and the longitudinal displacement prediction value, and the magnitude of the difference between the Doppler velocity prediction value and the Doppler velocity measurement value, then the state measurement value is identified as an outlier. The physical quantity to be compared whose difference magnitude exceeds the corresponding preset range is recorded as an outlier physical quantity to be compared, and the value of the outlier value discrimination counter corresponding to the outlier physical quantity to be compared is incremented by 1. The value is used to characterize the number of consecutive cumulative times of the outlier physical quantity to be compared.

[0086] In other words, it determines whether each of the three difference results falls within its respective preset range. Alternatively, this can be determined by comparing each difference result with the maximum threshold value of its respective preset range.

[0087] Suppose the maximum threshold corresponding to the physical quantity to be compared, or the discrimination threshold, is δ=[δ x ,δ y ,δ v ,], where δ x δ is the threshold for judging the magnitude of the difference between the measured and predicted lateral displacement values. y δ is the threshold for judging the magnitude of the difference between the measured and predicted longitudinal displacement values. v This is the discrimination threshold for the magnitude of the difference between the predicted and measured Doppler velocity values. The threshold value is obtained through comprehensive statistical analysis of radar measurement parameters and main application scenarios. Let the magnitude of the difference between the physical quantities to be compared be err = [err...]. x ,err y ,err v ,],err x err represents the magnitude of the difference between the measured and predicted lateral displacement values. y The magnitude of the difference between the measured longitudinal displacement and the predicted longitudinal displacement, err v Let be the magnitude of the difference between the predicted and measured Doppler velocity values, where:

[0088]

[0089] The magnitude of the difference between the physical quantities to be compared is compared with the discrimination threshold δ. If the magnitude of the difference between the physical quantities to be compared is greater than the discrimination threshold, that is, it exceeds the preset range, the state measurement value is considered to be a singular value, and this physical quantity to be compared is recorded as a singular physical quantity to be compared.

[0090] Furthermore, in one possible implementation disclosed in this invention, a singular value discrimination counter is set for each physical quantity to be compared. The value of the counter is used to characterize the consecutive cumulative number of times the physical quantity to be compared is recorded as a singular physical quantity to be compared. If, in the previous step, it is determined that the magnitude of the difference between the transverse displacement measurement value and the transverse displacement prediction value is greater than the discrimination threshold, and the magnitudes of the differences between the longitudinal displacement measurement value and the longitudinal displacement prediction value, and the Doppler velocity prediction value and the Doppler velocity measurement value are not greater than their respective discrimination thresholds, then the state measurement value is determined to be a singular value, the transverse displacement measurement value among the physical quantities to be compared is recorded as a singular physical quantity to be compared, and the value of its corresponding singular value discrimination counter is incremented by 1, i.e., the singular value counter N. x =N x +1. Meanwhile, the singular value discrimination counters for the longitudinal displacement and Leyne velocity, the physical quantities to be compared, remain unchanged.

[0091] If the magnitude of all differences is not greater than the discrimination threshold, then it is considered that the magnitude of all differences is within the preset range and the state measurement value is not a singular value. Then, the singular value counters of all physical quantities to be compared are set to 0, that is, all N=0.

[0092] In this embodiment, by setting an outlier counter, the number of times a state measurement value is identified as an outlier can be statistically analyzed. Furthermore, the count can be detailed down to the number of times each physical quantity to be compared is recorded as an outlier. In other words, the status of state measurements can be accurately recorded. If a state measurement value is continuously identified as an outlier, it's possible that the target's state has actually changed, and the state measurement value accurately reflects this change. However, continuing to identify it as an outlier in this case would lead to misjudgment and affect the actual detection accuracy. Therefore, the outlier counter helps technicians more intuitively identify this situation and adjust their strategies promptly.

[0093] In one possible implementation, in addition to recording the number of times the state measurement value is consecutively determined to be a singular value, it also records which physical quantity to be compared becomes a singular physical quantity to be compared, so as to intuitively show where the factors affecting the state measurement value become a singular value are.

[0094] When recording singular physical quantities to be compared, a flag array can be used. Taking the three physical quantities to be compared in this invention as an example, a 3*1 flag array flag=[0,0,0] can be pre-generated, representing whether the physical quantity to be compared is a singular physical quantity to be compared. For example, if it is 0, it is not a singular physical quantity to be compared; if it is 1, it means that the physical quantity to be compared is a singular physical quantity to be compared. Taking the transverse displacement measurement value as an example of being a singular physical quantity to be compared, the flag array flag=[1,0,0].

[0095] This embodiment does not limit the first position in the array to represent whether the lateral displacement measurement value is a singular physical quantity to be compared; it can be any one of the three positions. The key is that subsequent processing using the array can accurately locate and record the physical quantity to be compared as a singular physical quantity.

[0096] Combining the singular value discrimination counter and the tag array described above, a specific implementation process for correcting the state measurement value when it is a singular value is as follows: Figure 3 As shown, it includes:

[0097] Step S301: Determine the first correction coefficient of the corresponding singular physical quantity to be compared based on the value of each singular value discrimination counter.

[0098] For each singular physical quantity to be compared, a corresponding first correction coefficient β(N) is determined. As shown in the above embodiment, if the lateral displacement measurement value is recorded as a singular physical quantity to be compared, then the first correction coefficient β of the lateral displacement measurement value is determined. x (N). The first correction coefficient for other singular physical quantities to be compared follows the same principle. In this embodiment, the first correction coefficient is not a fixed value, but is related to the singular value discrimination counter value N. The purpose of the first correction coefficient is to multiply it with the corresponding singular physical quantity to be compared, and use the result of the multiplication as the corrected physical quantity to be compared, so that it is no longer a singular physical quantity to be compared, that is, the magnitude of the difference is 0, or to improve its singularity, that is, to reduce the magnitude of the difference.

[0099] Because N represents the number of times the state measurement value is consecutively judged as a singular value, if N is greater than the preset value, it means that the state measurement value is a singular value for a long period of time. This may be because the target's state has actually deviated significantly. In this case, the actual value has been misjudged as a singular value. Therefore, if it is still corrected as a singular value, it will affect the target's state detection results.

[0100] Therefore, in this step, the first correction factor for the corresponding singular physical quantity to be compared is set based on the value of the singular value discrimination counter. The setting method is as follows:

[0101] If the value of the singular value discrimination counter is not greater than a preset value, which can be a value set based on experience, such as 3 or 5, a first correction coefficient is set according to a first preset rule, which is to reduce the magnitude of the difference between the singular physical quantities to be compared after correction.

[0102] If the value of the singular value discrimination counter is greater than the preset value, then the first correction coefficient is set according to the second preset rule. The second preset rule is that the magnitude of the difference between the singular physical quantities to be compared remains unchanged or increases after correction.

[0103] When the value exceeds a preset value, it indicates that multiple consecutive measurements are singular values, suggesting a genuine shift in the target's state. Therefore, adjustments are no longer made to reduce the difference, but rather to align the measurement with its actual trajectory. Maintaining the magnitude of the difference means no adjustment is needed, and the first correction coefficient is 1, at least remaining unchanged. Increasing the correction coefficient means adjusting the measurement in the opposite direction to correct previous errors, resulting in a first correction coefficient greater than 1. Specific values ​​can be set empirically, such as 1.1 or 1.2. Alternatively, it can be set based on the first correction coefficient from the previous target detection process. For example, if the previous first correction coefficient was 0.6, meaning it was reduced by 40%, this value can be used as a reference for setting the first correction coefficient this time.

[0104] Step S302: Correct the measured values ​​of the singular physical quantities to be compared based on the first correction factor.

[0105] Taking the example of lateral displacement measurement being recorded as an unusual physical quantity to be compared, the correction process is expressed by the formula:

[0106] x = β·x m ,flag x =1 (Formula 2)

[0107] Step S303: Determine the correction value of the state measurement value based on the correction value of the singular physical quantity to be compared.

[0108] Based on the corrected lateral displacement measurement value, the corrected value of the state measurement value is obtained.

[0109] This embodiment sets a first correction coefficient based on the number of times singular values ​​occur, and then uses the first correction coefficient to correct the corresponding singular physical quantity to be compared. This method can promptly detect the situation of singular value misjudgment and correct the misjudgment result, thereby further improving the accuracy of target detection results.

[0110] In one possible implementation, combining the singular value discrimination counter and the tag array described above, another specific implementation process for correcting the state measurement value when it is a singular value is as follows: Figure 4 As shown, it includes:

[0111] Step S401: Determine the first correction coefficient and the second correction coefficient of the corresponding singular physical quantity to be compared based on the value of the singular value discrimination counter, wherein the sum of the first correction coefficient and the second correction coefficient is 1.

[0112] In this embodiment, two correction coefficients are determined for the singular physical quantity to be compared: a first correction coefficient β(N) and a second correction coefficient α(N). These two correction coefficients are used for correction. Furthermore, these two correction coefficients satisfy α(N) + β(N) = 1.

[0113] Step S402: Correct the measured value of the singular physical quantity to be compared based on the first correction coefficient, and correct the predicted value of the physical quantity to be compared based on the second correction coefficient.

[0114] Compared with the previous embodiment, in this embodiment, in addition to using the first correction coefficient to correct the measured value of the singular physical quantity to be compared, the second correction coefficient is also used to correct the predicted value of the physical quantity to be compared.

[0115] Step S403: Take the corrected predicted value of the singular physical quantity to be compared and the sum of the corrected predicted values ​​as the corrected value of the singular physical quantity to be compared.

[0116] After correcting both the predicted and measured values, their sum is used as the correction value for the singular physical quantity to be compared. In other words, the final correction value not only corrects the measured value of the singular physical quantity to be compared but also incorporates its predicted value. In this scheme, the predicted value serves as a key feature for determining whether a measured value is singular and possesses a certain degree of accuracy. Therefore, in this embodiment, introducing a more accurate predicted value to correct the measured value improves the accuracy of the correction result and further enhances the accuracy of the target detection result.

[0117] For example, assuming the lateral displacement measurement is recorded as a singular physical quantity to be compared, determine the first correction factor β for the lateral displacement measurement. x (N), second correction factor α x (N), the correction value x of the singular physical quantity to be compared in the flag array flag = [1,0,0] is obtained by the following formula:

[0118] x = α x ·x p +β x ·x m flag x =1 (Formula 3)

[0119] If the other two physical quantities to be compared are also recorded as singular physical quantities to be compared, then the flag array flag = [1,1,1] can be used to obtain their respective correction values ​​using the following formula:

[0120]

[0121] Each singular physical quantity to be compared is corrected using its own first correction factor and second correction factor to obtain its corresponding correction value.

[0122] Step S404: Determine the correction value of the state measurement value based on the correction value of the singular physical quantity to be compared.

[0123] In this embodiment, determining the first and second correction coefficients corresponding to the physical quantity to be compared based on the value of the singular value discrimination counter includes:

[0124] If the value of the singular value discrimination counter is not greater than a preset value, then the first correction coefficient and the second correction coefficient are set according to the first preset rule, wherein the first preset rule is to reduce the modulus of the difference between the physical quantities to be compared after correction.

[0125] If the value of the singularity discrimination counter is greater than a preset value, the first correction coefficient and the second correction coefficient are set according to the second preset rule, wherein the modulus of the difference between the physical quantities to be compared remains unchanged or increases after correction.

[0126] It can be seen that, with Figure 3 Compared to the illustrated embodiment, although the number of correction coefficients introduced has changed, the purpose of the correction is the same, so the principle for determining the correction coefficients is also similar, and the achieved correction effect is relative to... Figure 3 The illustrated embodiment has been further improved.

[0127] To further illustrate the effects of the embodiments of the present invention, an example of singular value identification and correction in the target detection process based on measured data is given below. For example... Figure 5 The figure shows the curve of the target's lateral position change during the test. The black point cloud data represents the measured values, the black box (i.e., Tracker1) represents the detection result obtained without using the target detection method of this invention, and the black diamond (i.e., Tracker2) represents the processing result obtained using the target detection method disclosed in the embodiments of this invention. As can be seen from the curve comparison in the figure, the detection result of Tracker1 is greatly affected by the singular values ​​in the measured values. It can be seen that around 660 frames, the result of Tracker1 shows obvious fluctuations. In contrast, the result of Tracker2 achieves a smooth transition of the detection result and is not significantly affected by the singular values ​​in the measured data. Therefore, the comparison shows that the detection method disclosed in this invention achieves real-time identification and correction of singular values ​​in the target detection process.

[0128] like Figure 6The figure shows the curve of the target's longitudinal position change during the test. Figure 6 It can be seen that there are obvious outliers in the measured values. Through comparative analysis, it can be seen that the detection method disclosed in this invention has achieved a good correction of the influence of outliers.

[0129] The present invention also discloses a target detection system, the structure of which is as follows: Figure 7 As shown, it includes:

[0130] The measurement value acquisition module 701 is used to acquire the state measurement value of the target at the current moment.

[0131] The prediction value acquisition module 702 is used to determine the state prediction value corresponding to the state measurement value.

[0132] The difference result calculation module 703 is used to calculate the difference result between the state prediction value and the state measurement value.

[0133] The singular value determination module 704 is used to determine whether a state measurement value is a singular value based on the difference results and preset discrimination rules.

[0134] The correction module 705 is used to correct the state measurement value when the state measurement value is an outlier.

[0135] The state correction value acquisition module 706 is used to obtain the state correction value of the target based on the corrected state measurement value.

[0136] The target detection system disclosed in this embodiment, after acquiring the target's state measurement value at the current moment, directly calculates the difference result using its corresponding state prediction value. Based on the calculation result, it determines whether the state measurement value is a singular value. If it is determined to be a singular value, the measurement value is corrected, and the corrected state value of the target is determined based on the correction result. Therefore, this invention can identify whether the measurement value is a singular value in real time and can correct the measurement value determined to be a singular value in real time, thereby greatly reducing the probability of inaccurate detection results due to the presence of singular values ​​and improving the continuity and stability of the track during target detection.

[0137] In one possible implementation, the difference result calculation module 703 includes:

[0138] The measurement value determination unit for the physical quantity to be compared is used to determine the measurement value of the physical quantity to be compared using the state measurement value; the physical quantity to be compared is used to describe the state of the target.

[0139] The predictive value determination unit for the physical quantity to be compared is used to determine the predicted value of the physical quantity to be compared using the state prediction value.

[0140] The calculation unit is used to calculate the difference between the measured value and the predicted value of the physical quantity to be compared, and uses the difference value as the difference result.

[0141] When the coordinate system of the state prediction value is a rectangular coordinate system and the coordinate system of the state measurement value is a polar coordinate system, the physical quantities to be compared include: Doppler velocity, longitudinal displacement and lateral displacement.

[0142] The calculation unit is specifically used to take the magnitude of the difference between the transverse displacement measurement value and the transverse displacement prediction value, the magnitude of the difference between the longitudinal displacement measurement value and the longitudinal displacement prediction value, and the magnitude of the difference between the Doppler velocity prediction value and the Doppler velocity measurement value as the difference results between the state measurement value and the state prediction value.

[0143] In one possible implementation, the singular value determination module 704 is specifically used to identify the state measurement value as a singular value when the difference result meets a preset discrimination rule. The preset discrimination rule includes: the difference result exceeds a preset range, and the preset range is the allowable range of the difference between the state measurement value and the state prediction value.

[0144] If the difference results meet the preset discrimination rules, the identified state measurement value is an outlier. This includes the following: at least one of the following exceeds its corresponding preset range: the magnitude of the difference between the lateral displacement measurement value and the lateral displacement prediction value, the magnitude of the difference between the longitudinal displacement measurement value and the longitudinal displacement prediction value, and the magnitude of the difference between the Doppler velocity prediction value and the Doppler velocity measurement value. If the state measurement value is identified as an outlier, the physical quantity to be compared whose difference magnitude exceeds the corresponding preset range is recorded as an outlier physical quantity to be compared. The value of the outlier discrimination counter corresponding to the outlier physical quantity to be compared is incremented by 1. The value is used to represent the number of consecutive cumulative times of the outlier physical quantity to be compared.

[0145] In one possible implementation, the correction module 705 includes:

[0146] The first correction coefficient determination unit is used to determine the first correction coefficient of the corresponding singular physical quantity to be compared based on the value of each singular value discrimination counter.

[0147] A measurement correction unit for a singular physical quantity to be compared is used to correct the measurement value of the singular physical quantity to be compared based on a first correction coefficient.

[0148] The correction value determination unit is used to determine the correction value of the state measurement value based on the correction value of the singular physical quantity to be compared.

[0149] The first correction coefficient determination unit determines the first correction coefficient of the corresponding singular physical quantity to be compared based on the value of each singular value discrimination counter, including:

[0150] If the value of the singular value discrimination counter is not greater than the preset value, then the first correction coefficient is set according to the first preset rule. The first preset rule is to reduce the modulus of the difference between the singular physical quantities to be compared after correction. Each singular physical quantity to be compared corresponds to a singular value discrimination counter.

[0151] If the value of the singular value discrimination counter is greater than the preset value, the first correction coefficient is set according to the second preset rule. The second preset rule is that the magnitude of the difference between the singular physical quantities to be compared remains unchanged or increases after correction.

[0152] In another possible implementation, the correction module 705 includes:

[0153] The first and second correction coefficient acquisition units are used to determine the first and second correction coefficients of the corresponding singular physical quantities to be compared based on the values ​​of the singular value discrimination counter, wherein the sum of the first and second correction coefficients is 1.

[0154] The correction unit is used to correct the measured value of the singular physical quantity to be compared based on the first correction coefficient, and to correct the predicted value of the physical quantity to be compared based on the second correction coefficient.

[0155] The summation calculation unit is used to take the corrected predicted value of the singular physical quantity to be compared and the sum of the corrected predicted values ​​as the corrected value of the singular physical quantity to be compared.

[0156] The correction value determination unit is used to determine the correction value of the state measurement value based on the correction value of the singular physical quantity to be compared.

[0157] The first and second correction coefficient acquisition units determine the corresponding first and second correction coefficients of the physical quantities to be compared based on the values ​​of the singular value discrimination counters. These correction coefficients include:

[0158] If the value of the singular value discrimination counter is not greater than the preset value, then the first correction coefficient and the second correction coefficient are set according to the first preset rule. The first preset rule is to reduce the magnitude of the difference between the physical quantities to be compared after correction.

[0159] If the value of the singular value discrimination counter is greater than the preset value, then the first correction coefficient and the second correction coefficient are set according to the second preset rule. The second preset rule is that the magnitude of the difference between the physical quantities to be compared remains unchanged or increases after correction.

[0160] The working process and principle of each functional module and unit in the target detection system disclosed in this embodiment can be referred to accordingly. Figures 1-6 Some of the methods described in the embodiments are not repeated here.

[0161] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0162] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0163] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0164] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A target detection method, characterized in that, include: Obtain the target's state measurement value at the current moment; Determine the predicted state value corresponding to the measured state value; Calculate the difference between the predicted state value and the measured state value; Based on the difference results and the preset discrimination rules, it is determined whether the state measurement value is an outlier. If the state measurement value is an outlier, the state measurement value is corrected. The state correction value of the target is obtained based on the corrected state measurement value; The correction of the state measurement value includes: The first correction coefficient of the corresponding singular physical quantity to be compared is determined based on the value of each singular value discrimination counter; The measured values ​​of the singular physical quantities to be compared are corrected based on the first correction factor; The correction value of the state measurement value is determined based on the correction value of the singular physical quantity to be compared.

2. The method according to claim 1, characterized in that, The result of calculating the difference between the predicted state value and the measured state value includes: The measured value of the physical quantity to be compared is determined using the state measurement value; the physical quantity to be compared is used to describe the state of the target. The predicted value of the physical quantity to be compared is determined using the state prediction value; Calculate the difference between the measured value and the predicted value of the physical quantity to be compared, and use the difference value as the difference result.

3. The method according to claim 2, characterized in that, When the coordinate system of the predicted state value is a rectangular coordinate system and the coordinate system of the measured state value is a polar coordinate system, the physical quantities to be compared include: Doppler velocity, longitudinal displacement, and lateral displacement. The step of calculating the difference between the measured value and the predicted value of the physical quantity to be compared, and using the difference as the difference result, includes: The magnitudes of the differences between the measured and predicted lateral displacement values, the magnitudes of the differences between the measured and predicted longitudinal displacement values, and the magnitudes of the differences between the predicted and measured Doppler velocity values ​​are used as the difference results between the measured and predicted state values.

4. The method according to claim 3, characterized in that, The process of determining whether the state measurement value is a singular value based on the difference result and the preset discrimination rule includes: If the difference result satisfies a preset discrimination rule, the state measurement value is identified as an outlier. The preset discrimination rule includes: the difference result exceeds a preset range, where the preset range is an allowable range representing the difference between the state measurement value and the state prediction value.

5. The method according to claim 4, characterized in that, The difference results satisfy a preset discrimination rule, and identifying the state measurement value as an outlier includes: If at least one of the following exceeds its corresponding preset range: the magnitude of the difference between the lateral displacement measurement value and the lateral displacement prediction value, the magnitude of the difference between the longitudinal displacement measurement value and the longitudinal displacement prediction value, and the magnitude of the difference between the Doppler velocity prediction value and the Doppler velocity measurement value, then the state measurement value is identified as an outlier. The physical quantity to be compared whose difference magnitude exceeds the corresponding preset range is recorded as an outlier physical quantity to be compared, and the value of the outlier value discrimination counter corresponding to the outlier physical quantity to be compared is incremented by 1. The value is used to characterize the number of consecutive cumulative times of the outlier physical quantity to be compared.

6. The method according to claim 1, characterized in that, The first correction coefficient for determining the corresponding singular physical quantity to be compared based on the value of each singular value discrimination counter includes: If the value of the singular value discrimination counter is not greater than a preset value, the first correction coefficient is set according to the first preset rule. The first preset rule is to reduce the modulus of the difference between the singular physical quantities to be compared after correction. Each singular physical quantity to be compared corresponds to a singular value discrimination counter. If the value of the singular value discrimination counter is greater than a preset value, the first correction coefficient is set according to the second preset rule, wherein the second preset rule is that the modulus of the difference between the singular physical quantities to be compared remains unchanged or increases after correction.

7. A target detection method, characterized in that, include: Obtain the target's state measurement value at the current moment; Determine the predicted state value corresponding to the measured state value; Calculate the difference between the predicted state value and the measured state value; Based on the difference results and the preset discrimination rules, it is determined whether the state measurement value is an outlier. If the state measurement value is an outlier, the state measurement value is corrected. The state correction value of the target is obtained based on the corrected state measurement value; The correction of the state measurement value includes: The first correction factor and the second correction factor are determined based on the value of the singular value discrimination counter, wherein the sum of the first correction factor and the second correction factor is 1; The measured value of the singular physical quantity to be compared is corrected based on the first correction coefficient, and the predicted value of the physical quantity to be compared is corrected based on the second correction coefficient; The sum of the corrected measured value and the corrected predicted value of the singular physical quantity to be compared is taken as the corrected value of the singular physical quantity to be compared. The correction value of the state measurement value is determined based on the correction value of the singular physical quantity to be compared.

8. The method according to claim 7, characterized in that, The result of calculating the difference between the predicted state value and the measured state value includes: The measured value of the physical quantity to be compared is determined using the state measurement value; the physical quantity to be compared is used to describe the state of the target. The predicted value of the physical quantity to be compared is determined using the state prediction value; Calculate the difference between the measured value and the predicted value of the physical quantity to be compared, and use the difference value as the difference result.

9. The method according to claim 8, characterized in that, When the coordinate system of the predicted state value is a rectangular coordinate system and the coordinate system of the measured state value is a polar coordinate system, the physical quantities to be compared include: Doppler velocity, longitudinal displacement, and lateral displacement. The step of calculating the difference between the measured value and the predicted value of the physical quantity to be compared, and using the difference as the difference result, includes: The magnitudes of the differences between the measured and predicted lateral displacement values, the magnitudes of the differences between the measured and predicted longitudinal displacement values, and the magnitudes of the differences between the predicted and measured Doppler velocity values ​​are used as the difference results between the measured and predicted state values.

10. The method according to claim 9, characterized in that, The process of determining whether the state measurement value is a singular value based on the difference result and the preset discrimination rule includes: If the difference result satisfies a preset discrimination rule, the state measurement value is identified as an outlier. The preset discrimination rule includes: the difference result exceeds a preset range, where the preset range is an allowable range representing the difference between the state measurement value and the state prediction value.

11. The method according to claim 10, characterized in that, The difference results satisfy a preset discrimination rule, and identifying the state measurement value as an outlier includes: If at least one of the following exceeds its corresponding preset range: the magnitude of the difference between the lateral displacement measurement value and the lateral displacement prediction value, the magnitude of the difference between the longitudinal displacement measurement value and the longitudinal displacement prediction value, and the magnitude of the difference between the Doppler velocity prediction value and the Doppler velocity measurement value, then the state measurement value is identified as an outlier. The physical quantity to be compared whose difference magnitude exceeds the corresponding preset range is recorded as an outlier physical quantity to be compared, and the value of the outlier value discrimination counter corresponding to the outlier physical quantity to be compared is incremented by 1. The value is used to characterize the number of consecutive cumulative times of the outlier physical quantity to be compared.

12. The method according to claim 7, characterized in that, The determination of the first and second correction coefficients corresponding to the physical quantity to be compared based on the value of the singular value discrimination counter includes: If the value of the singular value discrimination counter is not greater than a preset value, then the first correction coefficient and the second correction coefficient are set according to the first preset rule, wherein the first preset rule is to reduce the modulus of the difference between the physical quantities to be compared after correction. If the value of the singularity discrimination counter is greater than a preset value, the first correction coefficient and the second correction coefficient are set according to the second preset rule, wherein the modulus of the difference between the physical quantities to be compared remains unchanged or increases after correction.

13. A target detection system, characterized in that, include: The measurement value acquisition module is used to acquire the target's state measurement value at the current moment; The prediction value acquisition module is used to determine the state prediction value corresponding to the state measurement value; The difference result calculation module is used to calculate the difference result between the predicted state value and the measured state value; The singular value determination module is used to determine whether the state measurement value is a singular value based on the difference result and the preset discrimination rule; A correction module is used to correct the state measurement value when the state measurement value is an singular value. The state correction value acquisition module is used to obtain the state correction value of the target based on the corrected state measurement value; The correction module is specifically used for: The first correction coefficient of the corresponding singular physical quantity to be compared is determined based on the value of each singular value discrimination counter; The measured values ​​of the singular physical quantities to be compared are corrected based on the first correction factor; The correction value of the state measurement value is determined based on the correction value of the singular physical quantity to be compared.

14. A target detection system, characterized in that, include: The measurement value acquisition module is used to acquire the target's state measurement value at the current moment; The prediction value acquisition module is used to determine the state prediction value corresponding to the state measurement value; The difference result calculation module is used to calculate the difference result between the predicted state value and the measured state value; The singular value determination module is used to determine whether the state measurement value is a singular value based on the difference result and the preset discrimination rule; A correction module is used to correct the state measurement value when the state measurement value is an singular value. The state correction value acquisition module is used to obtain the state correction value of the target based on the corrected state measurement value; The correction module is specifically used for: The first correction factor and the second correction factor are determined based on the value of the singular value discrimination counter, wherein the sum of the first correction factor and the second correction factor is 1; The measured value of the singular physical quantity to be compared is corrected based on the first correction coefficient, and the predicted value of the physical quantity to be compared is corrected based on the second correction coefficient; The sum of the corrected measured value and the corrected predicted value of the singular physical quantity to be compared is taken as the corrected value of the singular physical quantity to be compared. The correction value of the state measurement value is determined based on the correction value of the singular physical quantity to be compared.

Images