CN106872971A - A kind of flying bird multiple targets tracking based on Swarm Intelligent Model - Google Patents

Abstract

The invention discloses a kind of flying bird multiple targets tracking based on Swarm Intelligent Model, flying bird multiple targets tracking proposed by the present invention is based on the flock of birds target data for obtaining in real time, the swarm intelligence in-flight embodied with flock of birds is as priori, consider the influence of the motion state to individual goal of adjacent objects, speed and positional information to individual goal are estimated, and combine the mass motion trend of multiple targets, predicted state to individual goal is modified, and realizes the tracking to each target in flock of birds.

Description

A kind of flying bird multiple targets tracking based on Swarm Intelligent Model

Technical field

The present invention relates to a kind of flying bird multiple targets tracking based on Swarm Intelligent Model, belong to low altitude airspace and supervise safely Depending on technical field, it is related to Radar Targets'Detection with tracking.

Background technology

The important technical that bird radar is the bird feelings observation of the Birds Nature Reserve such as estuarine wetland and research is visited, can be right Birds population scale and mechanics in protection zone make accurate count, assist relevant department to carry out the daily prison of protection zone birds Survey and research work, realize that whole day migratory bird feelings are monitored.

Bird feelings data processing is the core for visiting bird radar, and flock of birds target following is the important content of bird feelings data processing.Pass The flying bird multiple targets tracking of system only focuses on the mass motion of flock of birds, and the only center to flock of birds target is tracked, it is difficult to accurate Really count the quantity of flying bird target in flock of birds.

In fact, flock of birds remains able to cooperate with flight well in the case of no centralized Control, by between individuality Interaction keep " linear team " or " focus type team ".The flight of each individuality follows following basic principle in flock of birds： Individual speed is tried one's best, and individuality adjacent thereto is consistent, whole individualities avoid mutually touching between group center's aggregation, each individuality Hit, embody certain " swarm intelligence ".

The content of the invention

The invention aims to solve the above problems, propose a kind of flying bird multiple targets based on Swarm Intelligent Model with Track method, the method is applied to based on the flying bird multiple targets tracking for visiting bird radar, realizes the accurate statistics to flock of birds destination number.

A kind of flying bird multiple targets tracking based on swarm intelligence, comprises the following steps：

Step 1, individual goal state estimations；

Step 2, multiple targets state estimation；

Step 3, individual goal state revision.

The advantage of the invention is that：

Flying bird multiple targets tracking based on swarm intelligence can utilize " the colony's intelligence embodied in flock of birds collaboration flight Can ", the motion state of each target in tracking flock of birds, the destination number that accurate statistics flock of birds is included provides each target Movement locus.

Brief description of the drawings

Fig. 1 is the schematic diagram of the flying bird multiple targets tracking based on Swarm Intelligent Model of the invention；

Fig. 2 be the embodiment of the present invention image sequence in the 1st two field picture adjustment location schematic diagram；

Fig. 3 is the flying bird multiple targets tracking result schematic diagram of the embodiment of the present invention.

Specific embodiment

Below in conjunction with drawings and Examples, the present invention is described in further detail.

The present invention is a kind of flying bird multiple targets tracking based on Swarm Intelligent Model, as shown in figure 1, including following step Suddenly：

Step 1, individual goal state estimations；

If the flying bird multiple targets being made up of N number of flying bird target, it is combined into the location of k-1 moment all flying bird targets collectionWherein,Represent the position of k-1 moment targets i, i ∈ { 1,2 ..., N }, all flying bird mesh Target sets of speeds isWherein,Represent the speed of k-1 moment targets i.First, calculate The distance of all other target in each target i and group, and the s target closest with target i is picked out, it is designated as setMeet following condition

In formula,The position of k-1 moment targets j is represented,The speed of k-1 moment targets j is represented, θ is distance threshold.

At the k moment, the movement velocity of each target i is influenceed by s adjacent objects, its discreet valueCalculated by following formula

In formula, θ_iFor speed estimates adjustment factor.

Then the position of estimating residing for k moment targets i is：

In formula,It is the position of estimating of k moment targets i, Δ t represents the data update cycle of radar.

Step 2, multiple targets state estimation；

Position is estimated residing for the k moment all flying bird targets obtained based on step 1Calculate Position is estimated at multiple targets centerFor

Calculate the overall movement velocity of multiple targetsFor

Meanwhile, the flying bird multiple targets measurement information that the k moment obtains isWherein,Represent the k moment The adjustment location of target i, calculates the position that multiple targets measures centerFor：

Step 3, individual goal state revision；

The multiple targets state estimation result obtained using step 2, is entered to the flying bird individual goal speed discreet value in step 1 Row amendment, such as following formula

In formula,It is the erection rate of k moment targets i,It is speed amendment adjustment factor.

Then k moment target i present positions are modified to

In formula,It is the correction position of k moment targets i.

Embodiment：

The flying bird multiple targets tracking result based on radar data is proposed to the present invention in middle two-dimensional space below in conjunction with the accompanying drawings The flying bird multiple targets tracking based on Swarm Intelligent Model illustrated and described.

The present invention is a kind of flying bird multiple targets tracking based on Swarm Intelligent Model, is comprised the following steps：

Step 1, individual goal state estimations；

As shown in Fig. 2 radar image size be 480 × 480, the image lower left corner be the origin of coordinates, X-axis level to the right, Y-axis Vertically upward, a group flying bird target is along X-axis negative direction horizontal movement.Preceding scan cycle is labeled as k-1=0, Current Scan week Phase is labeled as k=1, the scan period Δ t=1 of radar.

In Fig. 2, the flying bird multiple targets being made up of N=35 flying bird target, residing for k-1=0 moment all flying bird targets Position beThe speed of all flying bird targets isAll targets with " " is marked in polar coordinate system.

By taking the target 1 in Fig. 2 as an example, first, the distance of target i=1 and all other target in group is calculated, and pick out The s=3 target closest with target i=1, is designated as setMeet following condition

In formula, θ=50 are distance threshold.The coordinate of target 1 is [252,356], the coordinate point of 3 targets adjacent thereto Wei not [239,356], [241,375] and [260,345]；The speed of target 1 is [- 9,0], the speed of 3 targets adjacent thereto Degree is respectively [- 8, -0.5], [- 9.5,0] and [- 10, -1].

At the k=1 moment, the movement velocity of each target i=1 is influenceed by s=3 adjacent objects, its discreet value Calculated by following formula

In formula,It is the movement velocity of k-1=0 moment target 1,It is the speed of target j in the adjacent objects set of target 1 Degree, θ₁For speed estimates adjustment factor.Based on formula (2), the movement velocity of target 1 is influenceed by neighbouring 3 targets, makes θ₁ =1, it is estimated speed and is

Then the position of estimating residing for k=1 moment target 1 is

In formula,Position is estimated for k=1 moment target 1,It is the position of k-1=0 moment target 1, Δ t=1 is represented The data update cycle of sensor.In this example, the position of estimating of target 1 is

Step 2, multiple targets state estimation；

Position is estimated residing for the k=1 moment all flying bird targets obtained based on step 1 Calculate multiple targets center estimates positionFor

Calculate the overall movement velocity of multiple targetsFor

Meanwhile, the flying bird multiple targets measurement information that the k=1 moment obtains isCalculate multiple targets amount The position of measured centerFor

Step 3, individual goal state revision；

The multiple targets state estimation result obtained using step 2, the speed to the flying bird individual goal 1 in step 1 is estimated Value is modified, such as following formula

In formula,It is the erection rate of k=1 moment target 1,It is speed amendment adjustment factor.OrderThen

Then the present position of k=1 moment target 1 is modified to

In formula,It is the correction position of k=1 moment target 1.

The state of k=1 moment all 35 targets is demarcated in Fig. 3, the correction position of each target has been represented by " ", by A piece short-term represents its revised direction of motion, and is added in satellite map.

Claims (4)

1. a kind of flying bird multiple targets tracking based on Swarm Intelligent Model, comprises the following steps：

Step one, individual goal state estimations；

Obtain k moment, the movement velocity discreet value of each target iEstimate position

Step 2, multiple targets state estimation；

The k moment is obtained, multiple targets measures the position at center

Step 3, individual goal state revision；

Obtain the correction position of k moment targets i.

2. a kind of flying bird multiple targets tracking based on Swarm Intelligent Model according to claim 1, described step One is specially：

If the flying bird multiple targets being made up of N number of flying bird target, it is combined into the location of k-1 moment all flying bird targets collectionWherein,Represent the position of k-1 moment targets i, i ∈ { 1,2 ..., N }, all flying birds The sets of speeds of target isWherein,Represent the speed of k-1 moment targets i；First, count The distance of each target i and all other target in group is calculated, and picks out the s target closest with target i, be designated as setMeet following condition

And j ≠ i (1)

In formula,The position of k-1 moment targets j is represented,The speed of k-1 moment targets j is represented, θ is distance threshold；

At the k moment, the movement velocity of each target i is influenceed by s adjacent objects, its discreet valueFor：

$v_{ip}^k=v_i^{k-1}+{\mathrm{\Σ}}_{(x_j^{k-1},v_j^{k-1})\∈S_i^{k-1}}\frac{{\mathrm{\θ}}_i}{|x_j^{k-1}-x_i^{k-1}|}\·(v_j^{k-1}-v_i^{k-1})---\left(2\right)$

In formula, θ_iFor speed estimates adjustment factor；

Then the position of estimating residing for k moment targets i is：

$x_{ip}^k=x_i^{k-1}+v_{ip}^k\·\mathrm{\Δ}t---\left(3\right)$

In formula,It is the position of estimating of k moment targets i, Δ t represents the data update cycle of radar.

3. a kind of flying bird multiple targets tracking based on Swarm Intelligent Model according to claim 1, described step Two are specially：

Position is estimated residing for the k moment all flying bird targets obtained based on step oneCalculate group Target's center estimates positionFor

$O_p^k=\frac{1}{N}\underset{i=1}{\overset{N}{\Σ}}{x}_{ip}^k---\left(4\right)$

Calculate the overall movement velocity of multiple targetsFor

$v_p^k=\frac{1}{N}\underset{i=1}{\overset{N}{\Σ}}{v}_{ip}^k---\left(5\right)$

Meanwhile, the flying bird multiple targets measurement information that the k moment obtains isWherein,Represent k moment targets The adjustment location of i, calculates the position that multiple targets measures centerFor：

$O_z^k=\frac{1}{N}\underset{i=1}{\overset{N}{\Σ}}{z}_i^k---\left(6\right).$

4. a kind of flying bird multiple targets tracking based on Swarm Intelligent Model according to claim 1, described step Three are specially：

The multiple targets state estimation result obtained using step 2, is carried out to the flying bird individual goal speed discreet value in step one Amendment, such as following formula

In formula,It is the erection rate of k moment targets i,It is speed amendment adjustment factor；

Then k moment target i present positions are modified to

$x_{iu}^k=x_i^{k-1}+v_{iu}^k\·\mathrm{\Δ}t---\left(8\right)$

In formula,It is the correction position of k moment targets i.