CN105223559B - A kind of long-range radar track initiation method that can switch parallel - Google Patents
A kind of long-range radar track initiation method that can switch parallel Download PDFInfo
- Publication number
- CN105223559B CN105223559B CN201510657048.4A CN201510657048A CN105223559B CN 105223559 B CN105223559 B CN 105223559B CN 201510657048 A CN201510657048 A CN 201510657048A CN 105223559 B CN105223559 B CN 105223559B
- Authority
- CN
- China
- Prior art keywords
- track
- beat
- theta
- hough transform
- flight path
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
Abstract
The invention discloses a kind of long-range radar track initiation method that can switch parallel, the amendment Hough transform track initiation method being delayed based on a step and quick amendment Hough transform track initiation method are run parallel, judgements of the discount judgement factor Dis to two kinds of Track initialization algorithms is wherein introduced to switch, and the overall merit to track initiation success rate, false track inception rate and initial time.With quick amendment Hough transform energy, quickly track initiation method, used time are shorter in the case of lower hybrid wave density.The amendment Hough transform based on step delay is used in the case of high clutter density, this method can preferably originate flight path in such a case.The present invention solves this by the weight matching perfection to principal element and when switches and most fast most could accurately originate flight path.
Description
Technical Field
The invention belongs to the field of radar signal processing, and particularly relates to a remote radar track starting method capable of realizing parallel switching.
Background
The track starting is used as the first step of tracking, and whether the fast track starting can be realized is obviously one of important factors whether the fast track can be realized. The track initiation exists under the condition of various track processing, the track initiation processing is relatively simple under the conditions of a single-target quiet environment and a multi-target quiet environment, and the track initiation processing is relatively complex under the conditions of a single-target noisy environment and a multi-target noisy environment. For multi-target track management, track initiation is the first step and is the basis for track tracking (maintenance), in recent years, with the emergence of complex new technologies, war environments become more and more complex, and a lot of research results also appear in the field of multi-target track management, but mainly aiming at the aspects of track fusion, target tracking and the like, the research results in the track initiation aspect are few. In addition, for engineering projects, a flight path starting method generally involving probability and likelihood function calculation can only be considered as a method discussion, and cannot be used in engineering projects. How to find an algorithm can find a better compromise between the fast track starting probability and the correct track starting probability, is convenient for engineering realization, and becomes the key point of research.
The standard Hough transform method belongs to a batch processing method, and simultaneously processes measured data of the last TotalScan scanning periods to determine a possible target track. The basic principle of Hough transformation is to transform a point in a measurement space to a curve or curved surface in a parameter space; points with the same parameter characteristics intersect in the transformed parameter space; and the characteristic curve detection is finished by judging the accumulation degree at the intersection point, so that whether the real track exists or not is judged. The Hough transformation has the advantages of insensitivity to local defect and robustness to random noise, is suitable for parallel processing and the like, and becomes a research hotspot of the current track initiation method. Random Hough Transform (RHT) -RHT is a probabilistic Hough transform proposed by Lei Xu et al. Compared to the standard Hough transform, the random Hough transform uses three new operation mechanisms: random sampling in image space, dynamic link list in parameter space, and convergence mapping connecting image space and parameter space. The RHT adopts one-to-one mapping, so that huge calculation amount caused by one-to-many mapping of standard Hough transformation is avoided; and a dynamic linked list structure is adopted, so that the memory requirement is reduced. RHT is widely used in the field of track initiation.
Modified Hough transform-an improvement proposed by Chen J on the basis of the standard Hough transform. Converting continuous TotalScan scanning time measurements received by a radar into a parameter space, and calculating a slope distance (parameter value) difference function; judging according to two initial criteria of the flight path:
(1) The angle of the zero crossing point must be very close.
(2) The slope sign of the angle across the zero crossing point must be the same.
If the conditions are met, whether the included angle between the target acceleration and the flight path meets certain constraints is judged, and if the included angle meets the constraints, the measurements can form a flight path. The method is characterized in that a condition is added on the basis of the modified Hough transformation, namely the modified Hough transformation can be used for transforming to a parameter space only when the measurement needs to meet the speed gating condition, so that the modified Hough transformation is transformed to the parameter space, the number of measuring points of the modified Hough transformation is greatly reduced, and the purpose of quickly starting a flight path is achieved.
However, for the corrected Hough transform track initiation algorithm based on one-step delay, the more the initial beat number TotalScan is, the longer the track initiation time is, the success rate of track initiation is not obviously improved, and the false track initiation rate is improved. The track starting success rate of the correction Hough transform track starting algorithm based on one-step delay has a good effect only when the detection probability Pd = 1. For the fast random Hough transform track initiation algorithm, the more the initial beat number TotalScan, the less the track initiation time is changed, the higher the track initiation success rate is, and the higher the false track initiation rate is. The larger the clutter density Lambda, the higher the false track inception rate.
Disclosure of Invention
The invention aims to provide a remote radar track starting method capable of realizing parallel switching, which overcomes the defects in the prior art.
In order to achieve the purpose, the invention adopts the following technical scheme:
a remote radar track starting method capable of switching in parallel comprises the following steps:
the method comprises the following steps: collecting two-dimensional energy data returned by the radar, calculating radar detection probability Pd, and calculating clutter density Lambda from the obtained TotalScan beat data;
step two: the discount judgment coefficient Dis is calculated by the following formula:
Dis＝0.2*3/TotalScan+0.6*Pd+0.2*Lambda/30；
wherein Dis represents a discount judgment coefficient, totalScan represents an initial beat number, pd represents radar detection probability, and Lambda represents clutter density;
step three: and when Dis is less than 0.8, a fast random Hough transform track starting method is adopted, and when Dis is more than or equal to 0.8, a modified Hough transform track starting method based on one-step time delay is adopted.
Further, the method also comprises the fourth step of: calculating a discount evaluation coefficient beta, and evaluating the track starting effect by adopting the discount evaluation coefficient beta, wherein the smaller the beta value is, the better the track starting effect is;
wherein, P T Is the track initiation success rate, P F Is the false track start rate and t is the start time.
Further, the fast random Hough transform track starting method comprises the following steps:
(1) Selecting a radar distance plane (x, y) as a measurement space, and defining a track parameter set as P = [ P ] c ,score]Initializing a track parameter set P = NULL, setting a threshold value T, wherein the sampling times gamma = 1;
(2) Defining data point D in TotalScan beat i ＝(x i ,y i ) Where i =1,2,3.. N, a measurement point z is randomly selected from the k beat i (k)＝(x i ,y i ) Randomly selecting a measuring point z from the k +1 th beat j (k+1)＝(x j ,y j )；
(3) Calculating the distance between two measurement points dist = | z i (k)-z j (k + 1) |, and the velocity v = dist/T between the two measurement points is determined s If, ifv min ≤v≤v max If not, entering the step (2); wherein, T s Showing the sampling interval of two measurement points, v max And v min Maximum and minimum speeds of the target, respectively;
(4) Calculating the straight line parameter p determined by two measuring points c ＝(θ 0 ,ρ 0 )；
(5) If k =1, turning to step (6); otherwise, find a P in P c ＝(θ c ,ρ c ) If | θ c -θ 0 Delta theta and rho are less than or equal to c -ρ 0 If | is less than or equal to Δ ρ, then p is c Score of (1) plus 1; otherwise p will be c Insert P and let P c Has a score of 1; where Δ θ and Δ ρ are tolerance errors;
(6) Limiting the sampling times by using a sampling termination rule, if the rule is met, finishing sampling at the k th beat and the k +1 th beat, and turning to the step (7) when k = k + 1; if the rule is not satisfied, turning to the step (2);
(7) If k is less than TotalScan, turning to the step (2); otherwise, P corresponding to score not less than T in P c Extracting and storing the parameters in a matrix Para, wherein the Para is the detected flight path parameters, and the flight path is started and ended.
Further, the modified Hough transform flight path starting method based on the one-step delay comprises the following steps:
(1) Defining data point D in TotalScan beat i ＝(x i ,y i ) Where i =1,2,3.. N, the maximum speed V of the target is set max And minimum velocity V min ；
(2) Taking the ith measurement point z of the kth beat of the radar i (k) As a starting point for the track start, a one-step extrapolation is made from the j-th metrology point z taken at k +1 j (k + 1) performing traversal correlation to obtain the velocity V between the k-th beat and the k + 1-th beat ij Will satisfy the condition V min ≤V ij ≤V max Storing the two measuring points;
(3) Taking the measuring points meeting the conditions in the step (2) as intermediate points of track initiation, performing one-step extrapolation, and taking the k +2 th shotm measurements z m (k + 2) performing ergodic correlation to obtain the speed V between the k +1 th beat and the k +2 th beat measuring point jm ；
(4) Solving the parameter theta of the k-th beat measuring point and the k + 1-th beat measuring point according to the formula rho = x cos theta + y sin theta ij And ρ ij And storing in a candidate flight path matrix P (k);
(5) Solving parameters theta of the k +1 th beat measuring point and the k +2 th beat measuring point according to the formula rho = x cos theta + y sin theta jm And ρ jm Respectively comparing with the parameters in P (k);
(6) If theta jm -θ ij | ≦ Δ θ, and θ ij And theta jm And if the slope is the same, wherein delta theta is an allowable error, the three shooting measuring points are simultaneously stored in a flight path matrix Para, the Para is a detected flight path parameter, and the flight path is started and ended.
Compared with the prior art, the invention has the following beneficial technical effects:
the method comprises the steps of correcting the Hough transform track starting method and quickly correcting the Hough transform track starting method based on one-step delay, running the Hough transform track starting method and the quickly corrected Hough transform track starting method in parallel, introducing judgment and switching of two track starting algorithms by using a discount judgment coefficient Dis, and applying the quickly corrected Hough transform quick track starting method under the condition of low clutter density, wherein the time consumption is short. And the method can better start the flight path in the environment by adopting one-step delay-based correction Hough transformation under the condition of high clutter density. The invention perfectly solves the problem that the fastest and most accurate initial track can be obtained only when switching is carried out by matching the weight of the main factors. The switching of the two algorithms realizes that the two algorithms can automatically arbitrate to realize the faster and more accurate initial flight path under the conditions of different clutter densities, different detection probabilities and different detection beats under the complex and changeable conditions. This is better than the existing single algorithm which can only adapt to the initial track in a specific environment.
Drawings
FIG. 1 is a flow chart of a track initiation method of the present invention;
fig. 2 is a graph (TotalScan = 3) of the modified Hough transform track start result based on one-step delay, wherein (a) represents the discount judgment coefficient values at different clutter densities; (b) A discount evaluation coefficient value representing different clutter densities is expressed; (c) The value of the track starting success rate under different clutter densities is represented; (d) Representing values of false track initiation rates at different clutter densities;
fig. 3 is a diagram of fast random Hough transform track start result (TotalScan = 3), where (a) represents discount judgment coefficient values under different clutter densities; (b) A discount evaluation coefficient value representing different clutter densities is expressed; (c) The value of the track starting success rate under different clutter densities is represented; (d) Representing values of false track initiation rates at different clutter densities;
fig. 4 is a graph of the modified Hough transform track initiation result (TotalScan = 10) based on one-step delay, wherein (a) represents the discount judgment coefficient values at different clutter densities; (b) A discount evaluation coefficient value representing different clutter densities is expressed; (c) The values of the track starting success rate under different clutter densities are represented; (d) Values representing false track initiation rates at different clutter densities;
fig. 5 is a graph of fast random Hough transform track initiation results (TotalScan = 10), wherein (a) represents discount decision coefficient values at different clutter densities; (b) A discount evaluation coefficient value representing different clutter densities is expressed; (c) The value of the track starting success rate under different clutter densities is represented; (d) Values representing false track initiation rates at different clutter densities.
Detailed Description
The invention is described in further detail below with reference to the accompanying drawings:
referring to fig. 1, since the track-start problem is different from static image processing, it is a dynamic process in which it appears most clearly that there is velocity information for either targets or clutter. The radar scanning image is not sampled for the whole measurement set, but sampled one by one, information is accumulated by using a sliding window method, data is processed in parallel by using Hough transformation, and track initiation is realized. Through research on a correction Hough transform track starting algorithm and a fast random Hough transform track starting algorithm based on one-step delay, the track starting success rate, the false track starting rate and the starting time of the two algorithms are closely related to the starting beat number TotalScan, the clutter density Lambda and the radar detection probability Pd. For the correction Hough transform track initiation algorithm based on one-step delay, the more the initial beat number TotalScan is, the longer the track initiation time is, the track initiation success rate is not obviously improved, and the false track initiation rate is improved. The track starting success rate of the modified Hough transform track starting algorithm based on one-step delay only has good effect when the detection probability Pd = 1. For the fast random Hough transform track initiation algorithm, the more the initial beat number TotalScan, the no obvious change in track initiation time, the higher the track initiation success rate and the higher the false track initiation rate. The clutter density Lambda is larger, the false track starting rate is higher, but the modified Hough transform track starting algorithm based on one-step delay is faster than the fast random Hough transform track starting algorithm. However, when the detection probability Pd ≠ 1, the algorithm still has a good track starting effect.
Aiming at the analysis, the invention introduces a discount judgment coefficient Dis and a discount evaluation coefficient beta. The method is used for judging and switching two track starting algorithms and comprehensively evaluating track starting success rate, false track starting rate and starting time. The discount judgment coefficient is calculated by adopting the following formula: dis =0.2 x 3/TotalScan +0.6 x Pd +0.2 x Lambda/30. And when Dis is more than or equal to 0.8, the algorithm is switched to correction Hough conversion based on one-step time delay, so that correct flight paths can be started well under high clutter density. When Dis is less than 0.8, the algorithm is switched to the fast correction Hough transformation, so that the initial flight path can be faster.
1) Random Hough transform track initiation algorithm
Is an arbitrary analytical curve in image space, wherein,are the n parameters of the curve. It can be appreciated that if parameters of a set of curves are obtained, a curve can be determined from them accordingly. Taking a straight line as an example, two points are randomly extracted from the binary edge image each time, and coordinate values are respectively obtained as (x) 1 ,y 1 ) And (x) 2 ,y 2 ) By solving the equation
A set of (p, θ) may be obtained and stored and accumulated in a dynamic join list. For each set of data, the dynamic join list consists of the parameter value p and the accumulation value score. Similarly, through several times of random sampling, the parameter value and the accumulation value in the parameter space are obtained, and the parameter corresponding to the accumulation value reaching the threshold is considered to be the parameter of the straight line in the image space.
The random Hough transform track starting algorithm is as follows:
inputting: data points D, D in all measurement spaces i ＝(x i ,y i )i＝1,2,3...n。
(1) The track parameter set is P = [ P = [ ] c ,score]Dynamic link list, initializing a track parameter set P = NULL, sampling times K =1, defining a maximum sampling time K max ；
(2) Two measurement points (x) were randomly selected from D i ,y i )，(x j ,y j )；
(3) Calculating a straight line parameter p = (theta) determined by two points 0 ,ρ 0 )；
(4) If k =1, turning to step (5); otherwise, find a P in P c ＝(θ c ,ρ c ) If | θ c -θ 0 Delta theta and rho are less than or equal to c -ρ 0 I ≦ Δ ρ (Δ θ and Δ ρ are tolerances, Δ θ = π/180 × 5, Δ ρ = 1), then p will be added c Score of (1) plus 1; otherwise, inserting P into P, and making score of P be 1;
(5) K = K +1, if K < K max Turning to the step (2), otherwise, completing sampling;
(6) P corresponding to that score is more than or equal to T in P c Extracted and stored in the matrix Para.
And (3) outputting: para is the detected flight path parameter.
2) Fast random Hough transform track initiation algorithm
Assuming that the target does uniform linear motion, selecting a radar range plane (x, y) as a measurement space, and sampling an interval T s . Let z i (k)＝(x i ,y i )，i＝1,...,l k Is the ith measurement trace at time k, the measurement itself contains velocity information. Different kinds of targets have their own velocity ranges, assuming maximum and minimum velocities, respectively, v max And v min . The transform function of the Hough transform is:
ρ＝x i cosθ+y i sinθ
first, the invalid samples are randomly sampled and reduced using the target motion information. Randomly extracting two points z from two adjacent beats in the measurement space i (k) And z j (k + 1). Calculate the distance between them dist = | z i (k)-z j (k + 1) |, and the velocity v = dist/T between two points is obtained s . If the speed limit v is satisfied min ≤v≤v max If so, the sampling is valid; and if the speed limit is not met, determining the sampling point as an invalid sampling point, and resampling until the speed limit is met.
And secondly, calculating and accumulating the flight path parameters. Solving the track parameter p between two points c = (= P, [ theta ]), and is stored in the parameter linked list P, while the accumulated value score of the parameter is recorded. If a certain calculated track parameter p c The same as the parameter value in P, only the corresponding accumulated value is updated, and the parameter does not need to be stored repeatedly.
Again, the number of samples is limited using a termination rule. And limiting the sampling times by using a sampling termination rule, finishing sampling if a certain condition is met, and otherwise, continuing sampling. And finally, extracting the flight path parameters. A threshold value T is set, and for all the accumulation values score in P,if score is more than or equal to T, extracting the corresponding parameter p c As detected track parameters.
The fast random Hough transform track starting algorithm comprises the following steps:
inputting: data points D, D in all measurement spaces (including TotalScan beat) i ＝(x i ,y i )i＝1,2,3...n。
(1) The track parameter set is P = [ P = [ ] c ,score]Initializing a track parameter set P = NULL, setting a threshold value T, and sampling times gamma = 1;
(2) Randomly selecting two measurement points z from D i (k)＝(x i ,y i ) And z j (k+1)＝(x j ,y j )；
(3) Calculating the distance between them dist = | z i (k)-z j (k + 1) |, and the velocity v = dist/T between two points is obtained s . If v is min ≤v≤v max Turning to the step (4); otherwise, turning to the step (2);
(4) Calculating a straight line parameter p determined by two points c ＝(θ 0 ,ρ 0 )；
(5) If k =1, turning to step (6); otherwise, find a P in P c ＝(θ c ,ρ c ) If | θ c -θ 0 Delta theta and rho are less than or equal to c -ρ 0 I ≦ Δ ρ (Δ θ and Δ ρ are tolerances, Δ θ = π/180 × 5, Δ ρ = 1), then p will be c Score of (1) plus 1; otherwise p will be c Insert P, let P c Has a score of 1;
(6) The number of samples is limited using a termination rule. If the rule is satisfied, ending sampling of the k th beat and the k +1 th beat, and turning to the step (7) when k = k + 1; if the rule is not satisfied, turning to the step (2);
the sampling termination rule is defined as follows:
assuming that the unknown parameter in the sample is θ, the parameter space is Θ. The hypothesis proposed as a precondition for the test is called the null hypothesis and is denoted as hypothesis H 0 ：θ∈Θ 0 . When a null hypothesis is rejected, logically meaning that a different hypothesis is accepted, this is calledFor the sake of alternative assumptions, it is denoted as assumption H 1 ：θ∈Θ 1 ，One is represented by H 0 Is a null hypothesis, H 1 The hypothesis testing problem for the alternative hypothesis is noted as:
whereinWhen theta is satisfied 1 ＝Θ-Θ 0 The alternative hypothesis is referred to as the logical opposite hypothesis of the zero hypothesis. If theta 0 ，Θ 1 Containing only one value, called H 0 ，H 1 Are in agreement with the assumption.
When the formula (1) is checked, two types of errors may be made. The first type of error is: when the zero hypothesis is set, the false reject zero hypothesis, which is called "false true rejection" because the sample falls in the reject domain, has a probability of P (D = H) 1 |H 0 ). The second type of error is: when the null hypothesis is not satisfied, the false accepted null hypothesis, called "false error", has a probability of P (D = H) because the sample falls in the accepted domain 0 |H 1 ). In hypothesis testing, this situation is typically met by increasing the probability of making a second type of error when trying to reduce the probability of making a first type of error; and vice versa. For this dilemma, the new man-pearson criterion, based on the principle of preserving the null hypothesis, is to specify a small positive number a, P in advance (D = H) 1 |H 0 ) Let P (D = H) under the condition of = alpha 0 |H 1 ) As small as possible. The accepted assumption is determined by equation (2):
wherein, X 1 ,X 2 ,...,X n Is sampleThe n samples in this space, λ, should be chosen to meet the requirement of a given positive number α.
The above is a description of the hypothesis testing problem with a fixed number of samples. In many cases, it is possible to obtain a better impression if the observation time or the number of samples is not determined before the observation, but the question when to terminate the observation is decided during the observation according to the situation of the result sought. Sequence detection refers to hypothesis testing in which the number of observation samples or observation time is not fixed when making a decision, that is, the decision made in each observation is not two but three, that is, a null hypothesis is accepted, an alternative hypothesis is accepted, and the observation is continued, which is watt sequence detection. The decision made is described by equation (3):
wherein A, B are each according to the specified P (D = H) 0 |H 1 ) And P (D = H) 1 |H 0 ) The calculated threshold value.
In the random Hough transform track starting algorithm, the determination of the optimal sampling number can be realized through sequence detection. Suppose that a total of k different track parameter values, X, were recorded in the dynamic link list during the first n random samples j Denotes the jth track parameter, X j ∈{X 1 ,X 2 ,...,X k }；X j Corresponding track parameter accumulation value is n j The whole random sampling process should obey the following three rules:
(l) The detection task of the flight path parameters is completed by using as few random samples as possible;
(2) The number of random samples should be greater than or equal to 2;
(3) The jth track parameter X in the dynamic link list j Can be described as P (X = X) j )＝p j ,j＝1,2,...,k。j max ＝arg max n j Representing the track parameter with the greatest accumulated valueThe sequence number in the dynamic join list, then the probability of its corresponding track parameter obeys the distribution.
In the first n random samples, ifProbability of (2)Andprobability P of j Satisfy a certain relationship, thenCan be considered as the final track parameter. At this point, the sampling may be terminated. Due to the fact thatTo the whole bodyThe comparison is not practical and further complicates the problem, so that only the parameter probabilities corresponding to the largest and second largest accumulation values need to be compared.
The following assumptions were made:
H 0 parameter corresponding to maximum accumulation value in dynamic link listIs the final starting track parameter.
H 1 Parameter corresponding to next largest accumulation value in dynamic link listIs the final starting track parameter.
The random Hough transform must be performed according to the sampling rule of equation (3) when sampling. If the error probability given in advance is P (D = H) 1 |H 0 )＝α，P(D＝H 0 |H 1 ) By derivation, = β, the relationship of α, β and a, B can be found as follows:
according to a guard null hypothesis H 0 The selection of α, β must satisfy the condition A>1>B>, 0. According to the above description, the sampling termination rule of the random Hough transform track start is composed of the formula (3), the formula (4) and the formula (5).
(7) If k is less than TotalScan, turning to the step (2); otherwise, P corresponding to Score more than or equal to T in P c And extracting and storing in a matrix Para, and ending the track start.
And outputting that Para is the detected flight path parameter.
3) Correction Hough transform track initial algorithm based on one-step time delay
The algorithm can overcome the defects that the one-step delay track starting algorithm cannot be applied to the dense clutter environment, the traversal calculation amount of the modified Hough transform track starting algorithm is large, and the implementation is not facilitated, and the track starting is realized quickly and efficiently. The algorithm adopts an improved multi-hypothesis algorithm to process the associated data, simultaneously adopts a sliding window method and multiple filtering to accumulate the flight path information, and finally determines the flight path of the target by using a Hough transformation algorithm.
The algorithm comprises the following steps:
inputting: data points (x) in all measurement spaces i ,y i )，x i And y i Representing measured position information, i =1,2,3.. N.
(1) Setting the maximum speed V max And minimum velocity V min The following pairThe three continuous beat measurements are processed.
(2) Taking the ith measurement z of the kth beat of the radar i (k) As a starting point for the track start, a one-step extrapolation is performed with the jth measurement z taken at the k +1 th j (k + 1) performing traversal correlation to obtain the velocity V between the k-th beat and the k + 1-th beat ij . Will satisfy the condition V min ≤V ij ≤V max The two beats are measured and stored.
(3) Taking the measurement satisfying the condition (2) as the intermediate point of the track start, performing a one-step extrapolation, with the m-th measurement z taken at k +2 m (k + 2) performing traversal correlation to obtain the velocity V between the measurement of the k +1 th beat and the measurement of the k +2 th beat jm 。
(4) Solving the parameter theta of the k-th beat measurement and the k + 1-th beat measurement according to the formula rho = x cos theta + y sin theta ij ，ρ ij Stored in the candidate track matrix P (k).
(5) Solving the parameter theta of the k +1 st beat measurement and the k +2 th beat measurement according to the formula rho = x cos theta + y sin theta jm ，ρ jm Respectively, with the parameters in P (k), and if the criterion one is satisfied: [ theta ] jm -θ ij And | is less than or equal to delta theta and criterion two: theta ij And theta jm The slope of the position is the same, wherein delta theta is an allowable error, the three-shot measurement is simultaneously stored in a track matrix Para, and the result is an initial result of the three-shot measurement;
and (3) outputting: the stored in Para is the initial track, and the track is completely started.
Fig. 1 is a flowchart of the algorithm, from which it can be seen visually how the present invention switches between two algorithms by the judgment of the discount judgment coefficient Dis, where the discount evaluation coefficient β is a discount factor generally smaller than 1, and as can be seen from fig. 2 to fig. 5, the higher the track initiation success rate is, the lower the false track initiation success rate is, the shorter the initiation time is, and the better the track initiation effect is. The lower the discount evaluation coefficient, the better the track initiation effect. When Dis is more than or equal to 0.8, the discount evaluation coefficient of the correction Hough transformation track starting algorithm based on one-step delay is lower than that of the quick correction Hough transformation track starting algorithm, and the track starting effect is good; when Dis is less than 0.8, the fast random Hough transform track starting algorithm has a lower discount evaluation coefficient than a modified Hough transform track starting algorithm based on one-step delay, and the track starting effect is good.
Claims (4)
1. A remote radar track starting method capable of being switched in parallel is characterized by comprising the following steps:
the method comprises the following steps: collecting two-dimensional energy data returned by a radar, calculating the detection probability Pd of the radar, and calculating clutter density Lambda from the obtained TotalScan beat data;
step two: the discount judgment coefficient Dis is calculated by the following formula:
Dis＝0.2*3/TotalScan+0.6*Pd+0.2*Lambda/30；
wherein Dis represents a discount judgment coefficient, totalScan represents an initial beat number, pd represents radar detection probability, and Lambda represents clutter density;
step three: and when Dis is less than 0.8, adopting a rapid random Hough transform track starting method, and when Dis is more than or equal to 0.8, adopting a one-step delay-based correction Hough transform track starting method.
2. The method for remote radar track initiation capable of being switched in parallel according to claim 1, further comprising the following four steps: calculating a discount evaluation coefficient beta, and evaluating the track starting effect by adopting the discount evaluation coefficient beta, wherein the smaller the beta value is, the better the track starting effect is;
wherein, P T Is the track initiation success rate, P F Is the false track start rate and t is the start time.
3. The remote radar track starting method capable of being switched in parallel according to claim 1, wherein the fast random Hough transform track starting method comprises the following steps:
(1) Selecting a radar distanceThe plane (x, y) is used as a measurement space, and the parameter set for defining the flight path is P = [ P ] c ,score]Initializing a track parameter set P = NULL, setting a threshold value T, wherein the sampling times gamma = 1;
(2) Defining data point D in TotalScan beat i ＝(x i ,y i ) N, wherein i =1,2,3.. N, a measurement point z is randomly selected from the k beat i (k)＝(x i ,y i ) Randomly selecting a measuring point z from the k +1 th beat j (k+1)＝(x j ,y j )；
(3) Calculating the distance between two measurement points dist = | z i (k)-z j (k + 1) |, and the velocity v = dist/T between the two measurement points is determined s If v is min ≤v≤v max If not, entering the step (2); wherein, T s Representing the sampling interval of two measurement points, v max And v min Maximum and minimum speeds of the target, respectively;
(4) Calculating the straight line parameter p determined by two measuring points c ＝(θ 0 ,ρ 0 )；
(5) If k =1, turning to step (6); otherwise, find a P in P c ＝(θ c ,ρ c ) If | θ c -θ 0 Delta theta and rho are less than or equal to c -ρ 0 If | is less than or equal to Δ ρ, then p is c Plus 1; otherwise p will be c Inserting P and let P c Has a score of 1; wherein Δ θ and Δ ρ are tolerance errors;
(6) Limiting the sampling times by using a sampling termination rule, if the rule is met, finishing sampling on the k th beat and the k +1 th beat, and turning to the step (7) when k = k + 1; if the rule is not satisfied, turning to the step (2);
(7) If k is less than TotalScan, turning to the step (2); otherwise, P corresponding to score not less than T in P c Extracting and storing the parameters in a matrix Para, wherein the Para is the detected flight path parameters, and the flight path is started and ended.
4. The remote radar track starting method capable of being switched in parallel according to claim 1, wherein the modified Hough transform track starting method based on one-step time delay comprises the following steps of:
(1) Defining data point D in TotalScan beat i ＝(x i ,y i ) Where i =1,2,3.. N, the maximum speed V of the target is set max And a minimum velocity V min ；
(2) Taking the ith measuring point z of the kth beat of the radar i (k) As a starting point for the track start, a one-step extrapolation is performed with the j-th measurement point z of the (k + 1) -th shot j (k + 1) performing ergodic correlation to obtain the velocity V between the k-th beat and the k + 1-th beat ij Will satisfy the condition V min ≤V ij ≤V max Storing the two measuring points;
(3) Taking the measuring points meeting the conditions in the step (2) as intermediate points of the track start, performing one-step extrapolation, and taking the mth measuring z of the k +2 beat m (k + 2) performing ergodic correlation to obtain the velocity V between the measuring points of the (k + 1) th beat and the (k + 2) th beat jm ；
(4) Solving the parameter theta of the k-th beat measuring point and the k + 1-th beat measuring point according to the formula rho = xcos theta + ysin theta ij And ρ ij And storing in a candidate flight path matrix P (k);
(5) Solving parameters theta of the k +1 th beat measuring point and the k +2 th beat measuring point according to the formula rho = xcos theta + ysin theta jm And ρ jm Respectively comparing with the parameters in P (k);
(6) If theta jm -θ ij | < delta theta, and theta ij And theta jm And if the slope is the same, wherein delta theta is an allowable error, the three shooting measuring points are simultaneously stored in a flight path matrix Para, the Para is a detected flight path parameter, and the flight path is started and ended.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510657048.4A CN105223559B (en) | 2015-10-13 | 2015-10-13 | A kind of long-range radar track initiation method that can switch parallel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510657048.4A CN105223559B (en) | 2015-10-13 | 2015-10-13 | A kind of long-range radar track initiation method that can switch parallel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105223559A CN105223559A (en) | 2016-01-06 |
| CN105223559B true CN105223559B (en) | 2018-02-27 |
Family
ID=54992617
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510657048.4A Active CN105223559B (en) | 2015-10-13 | 2015-10-13 | A kind of long-range radar track initiation method that can switch parallel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105223559B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108919268B (en) * | 2018-06-29 | 2020-11-24 | 安徽四创电子股份有限公司 | Track initiation algorithm based on unmanned aerial vehicle monitoring radar |
| CN110554365A (en) * | 2019-09-02 | 2019-12-10 | 北京电子工程总体研究所 | Seeker multi-frequency point difference directional diagram automatic calibration method and device |
| CN111781592B (en) * | 2020-06-12 | 2023-12-22 | 中国船舶集团有限公司第七二四研究所 | Rapid automatic starting method based on fine granularity characteristic analysis |
| CN113671482B (en) * | 2021-09-18 | 2022-04-08 | 哈尔滨工业大学 | Track initiation method based on multi-dimensional screening parallel Hough transformation |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6826311B2 (en) * | 2001-01-04 | 2004-11-30 | Microsoft Corporation | Hough transform supporting methods and arrangements |
| US7406212B2 (en) * | 2005-06-02 | 2008-07-29 | Motorola, Inc. | Method and system for parallel processing of Hough transform computations |
| CN103116160A (en) * | 2012-11-27 | 2013-05-22 | 电子科技大学 | Weak target detecting method based on random Hough transformation |
| CN104569923B (en) * | 2015-02-05 | 2017-05-03 | 中国人民解放军海军航空工程学院 | Velocity restraint-based Hough transformation fast track starting method |
-
2015
- 2015-10-13 CN CN201510657048.4A patent/CN105223559B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105223559A (en) | 2016-01-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105223559B (en) | A kind of long-range radar track initiation method that can switch parallel | |
| CN108490410B (en) | Two-coordinate radar sea target joint detection and tracking method | |
| CN107728140B (en) | Multi-target multi-channel parallel tracking processing method for warning radar | |
| CN108645412B (en) | Multi-sensor self-adaptive track starting method | |
| CN109858526B (en) | Multi-target track fusion method based on sensor in target tracking | |
| CN107202989B (en) | Complex weak target detection and tracking method suitable for passive towed linear array sonar | |
| CN109613483B (en) | Multi-target track initiation method based on Hough transformation | |
| CN112130142A (en) | Micro Doppler feature extraction method and system for complex moving target | |
| CN110849372A (en) | Underwater multi-target track association method based on EM clustering | |
| CN111123205A (en) | Ground reconnaissance radar target point trace condensing method | |
| CN109190647B (en) | Active and passive data fusion method | |
| CN110488273A (en) | A kind of vehicle tracking detection method and device based on radar | |
| CN111366900A (en) | Tracking radar track quality evaluation method, system and medium based on residual error statistics | |
| CN110045363B (en) | Multi-radar track association method based on relative entropy | |
| CN108828584B (en) | Multi-frequency target tracking-before-detection method based on track folding factor ambiguity resolution | |
| CN115220002A (en) | Multi-target data association tracking method and related device for fixed single station | |
| Bi et al. | Improved multi-target radar TBD algorithm | |
| Silantyev et al. | Track Association in Radars with Additional Multi-Scanning Observation Channels | |
| Timoshenko et al. | Detection and estimation of parameters of a random process set in multi-Scanning radar observation based on the" track-before-detect" methods | |
| CN114114242A (en) | Radar target tracking method based on point-selected flight path and flight path extrapolation | |
| CN113534135A (en) | Flight path correlation method and device based on dispersion linear trend test | |
| Wang et al. | Research on multi-maneuvering target tracking JPDA algorithm | |
| Ge et al. | Multi-radar hybrid detection algorithm based on information entropy | |
| CN113192110A (en) | Multi-target tracking method, device, equipment and storage medium | |
| CN113408422B (en) | Multi-frame joint detection tracking and classification method suitable for weak targets |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210201 Address after: Room 508, block a, Rongcheng cloud Valley, 57 Keji 3rd road, Zhangba Street office, high tech Zone, Xi'an City, Shaanxi Province, 710075 Patentee after: Xi'an Huizhi Information Technology Co.,Ltd. Address before: 710064 No. 33, South Second Ring Road, Shaanxi, Xi'an Patentee before: CHANG'AN University |
|
| TR01 | Transfer of patent right |