CN103369623A - System and method for dynamically switching target tracking method in sensor network - Google Patents

Abstract

The invention provides a method for dynamically switching a target tracking method in a sensor network. The method comprises the following steps: first, an information fusion node estimates an error between a predicted value and an actual measured value of a next position of a moving target based on target state information from a target tracking node so as to obtain a dynamic switching reference value dsr and sends a message to a cooperative node when the dsr is greater than a predetermined threshold value omega; second, the cooperative node judges the current number of targets according to an image, acquired through an image sensor, of a current scene and sends the number to the information fusion node as a judgment result; and finally, the information fusion node judges whether a target tracking method needs to be switched currently according to the dynamic switching reference value and the judgment result from the cooperative node. The method makes full use of the advantages of a chip sensor node and an image sensor node and can be used for finding a timing for target switching more accurately and more quickly.

Description

Dynamically switch the system and method for method for tracking target in the sensor network

Technical field

The invention belongs to the target tracking domain in the sensor network, relate in particular to the switching of carrying out method for tracking target.

Background technology

In sensor network, often need the target by transducer detection and tracking campaign.Existing tracking to moving target comprises monotrack method and multi-object tracking method.Effect when these two kinds of methods are followed the tracks of target is different, if go to follow the tracks of multiple target with the monotrack method, the effect of target following can be very poor; Equally, if go the tracking report target with multi-object tracking method, the effect of target following also can be very poor.Present track algorithm is assumed to monotrack or multiple target tracking mostly in advance, and target numbers is fixing in advance, therefore can select in advance and adopt monotrack algorithm or multiple target tracking algorithm.

But in Sensor Network target following scene, dynamic change often occurs in destination number, wants target is followed the tracks of accurately, will use exactly corresponding tracking, and opportunity that so how to find method for tracking target to switch is just very important.In sensor network, also do not have at present technology to change according to target numbers and carry out the dynamic switching of method for tracking target.

Summary of the invention

Therefore, the object of the invention is to solve the problems of the technologies described above, a kind of method of in sensor network method for tracking target dynamically being switched is provided, the opportunity that can find accurately and rapidly single goal and multiple target to occur.

The objective of the invention is to be achieved through the following technical solutions:

On the one hand, the invention provides a kind of method that method for tracking target is dynamically switched for sensor network, described method comprises:

Step 1) predicted value of the next position of moving target and the error between the actual measured value are estimated based on coming from the target status information of target following node by the information fusion node, dynamically to be switched reference value dsr;

Step 2) as resulting dynamic switching reference value dsr during greater than predetermined threshold value ω, sends message to cooperative nodes, determine current target number with the indication cooperative nodes;

Step 3) by cooperative nodes after the message of receiving from the information fusion node, judge current target number according to the image of current scene, and it sent to the information fusion node as judged result;

Step 4) by the information fusion node according to dynamically switching reference value and from the judged result of cooperative nodes, judging the current method for tracking target that whether needs to switch.

In the said method, described cooperative nodes can be imageing sensor.

In the said method, in step 1) described dynamic switching reference value dsr can for: ζ wherein _iBe i the component of vectorial d (k), d (k)=(ζ ₁, ζ ₂..., ζ _n) ^TThat the k moment is to the predicted value of the next position of moving target and the estimation of the error between the measured value.

In the said method, in step 2) can also comprise if dsr, calculates based target greater than predetermined threshold value ω and follow the tracks of the confidence values φ that the information of node is judged the dynamic switching time of method for tracking target ₁Step, φ wherein ₁Computing formula be

0＜φ wherein ₁＜1.

In the said method, in step 3) can also comprise and calculate the confidence values φ that cooperative nodes is judged the dynamic switching time of method for tracking target ₂, and send it to the step of information fusion node, wherein φ ₂Computing formula is

0＜φ wherein ₂＜1, u=1,2 ... p, p represent the target number that current time detects, and q represents original target number, S _uThe area that represents the pixel region of u target, σ are the threshold values of being scheduled to.

In the said method, in step 4) can calculate confidence values δ with following formula by the information fusion node:

δ＝φ ₁+φ ₂

When the value of δ during greater than threshold value ψ, indicating target is followed the tracks of node and is switched to corresponding target tracking algorism and carry out target following.

In the said method, threshold value ψ can be according to different scene spans between 1.2～1.8.

In the said method, in step 4) can also comprise: as dynamic switching reference value dsr during greater than predetermined threshold value ω, if be single target from the current target number of cooperative nodes, then indicating target is followed the tracks of node and is switched to the monotrack algorithm; If the current target number from cooperative nodes is a plurality of targets, then indicating target is followed the tracks of node and is switched to the multiple target tracking algorithm.

Another aspect the invention provides a kind of system that method for tracking target is dynamically switched for sensor network, and described system comprises:

The target following node is used for the pursuit movement target, and sends target status information to the information fusion node;

The information fusion node is used for based on coming from the target status information of target following node to the estimating of error between the predicted value of the next position of moving target and the actual measured value, dynamically to be switched reference value dsr; And as resulting dynamic switching reference value dsr during greater than predetermined threshold value ω, send message to cooperative nodes, determine current target number with the indication cooperative nodes; And be used for according to dynamically switching reference value and from the judged result of cooperative nodes, judging the current method for tracking target that whether needs to switch;

Cooperative nodes is used for judging current target number according to the image of current scene, and it being sent to the information fusion node as judged result after the message of receiving from the information fusion node.

Wherein, described cooperative nodes can be imageing sensor.

Compared with prior art, the invention has the advantages that:

Judge whether to carry out the switching of method for tracking target in conjunction with the information from target following node and cooperative nodes.Serve as the target following node by the sensor node of cheapness target is carried out real-time tracking.Serve as cooperative nodes by the imageing sensor node, only in needs, carry out the detection of destination number, take full advantage of the advantage of cheap sensor node and imageing sensor node, the opportunity that can find more accurate, faster target to switch.

Description of drawings

Embodiments of the present invention is further illustrated referring to accompanying drawing, wherein:

Fig. 1 is the schematic flow sheet according to the method that method for tracking target is dynamically switched of the embodiment of the invention.

Fig. 2 (a) and Fig. 2 (b) are the Application Scenarios-Example schematic diagram according to the embodiment of the invention.

Embodiment

In order to make purpose of the present invention, technical scheme and advantage are clearer, and the present invention is described in more detail by specific embodiment below in conjunction with accompanying drawing.Should be appreciated that specific embodiment described herein only in order to explain the present invention, is not intended to limit the present invention.

Fig. 1 has provided the schematic flow sheet according to the method that method for tracking target is dynamically switched of the embodiment of the invention.The method is followed the tracks of the information of node collection and the information that gathers based on cooperative nodes in conjunction with based target, judges comprehensively whether the dynamic switching time of method for tracking target arrives.

The below is take the concrete steps of typical application scenarios method shown in Figure 1 as example illustrates of the wireless sensor network shown in Fig. 2 (a).Shown in Fig. 2 (a), this application scenarios is the vehicle tracking in the intelligent transportation system.With some sensor nodes and the imageing sensor node deployment both sides at road.These nodes can be divided into a group every a segment distance, comprise sensor node and imageing sensor node in each group, these cheap sensor nodes intercom mutually with a small amount of imageing sensor node, form a kind of heterogeneous Sensor Network.Wherein, sensor node serves as the target detection node, is used for checking the appearance of target; And after detecting target, these sensor nodes can serve as the target following node again, are used for target is followed the tracks of.In other embodiments, also can be that wherein a part of sensor node serves as the target detection node, be used for checking the appearance of target; And another part sensor node serves as the target following node, is used for target is followed the tracks of.The imageing sensor node serves as cooperative nodes, is used for current goal is carried out IMAQ and judges the number of target.And, also comprising the information fusion node in this system, the information that is used for information that comprehensive all target following nodes gather and cooperative nodes judges whether to occur target and switches.Can serve as described information fusion node by one of them of target following node or the sensor node of special use.In addition, can also comprise a clock synchronous node, to be used for that each internodal clock is consistent.

During initial condition, the target detection node is in active state, and information fusion node and cooperative nodes all are in sleep state.When detecting target and occur, the target detection node will serve as the target following node to be followed the tracks of target, wakes simultaneously the information fusion node up; The information that gathers based on the target following node when the information fusion node is judged and may be occurred that target is switched or when abnormal conditions occurring, wake cooperative nodes up; The result that the information that cooperative nodes gathers according to self is judged and will be judged issues the information fusion node; The judged result that the information of the comprehensive based target tracking of information fusion node node is done and the judged result of cooperative nodes are made final judgement, if target occurs to be switched, then switch to corresponding target tracking algorism and carry out target following, otherwise be left intact.

Also show the member in certain concrete group in the application scenarios shown in Fig. 2 (a) among Fig. 2 (b), for example comprise four target following nodes (sensor node 10, sensor node 11, sensor node 20, sensor node 21, their also are the target detection nodes simultaneously), a clock synchronous node 51 (being used for each internodal clock is consistent), an information fusion node (sensor node 50), two cooperative nodes ( imageing sensor node 52 and 53).Wherein sensor node 10,11,20,21 these four target following nodes will collect target information and carry out simple feature calculation at node, then the target status information after will calculating for example passes to information fusion node 50 by radio frequency, carry out information fusion at this information fusion node 50, obtain target in this state vector constantly, then by the estimation to error between the predicted value of the next position of moving target and the measured value, judge that the current target that whether occurs switches or occur unusual.For example, estimated error can be weighed by dynamic switching reference value dsr, if the value of dsr has exceeded given threshold range, just think and the target switching occurs or occur unusual, at this moment information fusion node 50 will send message to imageing sensor node 52 and 53, so that imageing sensor is started working, gather the image of current goal, judge this moment target number, and this judged result also sent to the information fusion node, information fusion node 50 comprehensive based targets are followed the tracks of judged result that the information of node collections makes and the judged result of cooperative nodes is made final judgement.When judging the needs switching, indicating target is followed the tracks of node and is switched tracking.

More specifically, the method mainly comprises the steps:

Step 1 is gone out target status information and sends it to the information fusion node according to the information calculations that self gathers by each target following node.Wherein target status information comprises position, speed, acceleration etc., is represented by the vectorial X in the following formula, and this step can adopt existing technology to finish.

Step 2 is estimated the predicted value of the next position of moving target and the error between the actual measured value based on coming from the target status information of target following node by the information fusion node, dynamically to be switched reference value.

In fact, when moving target is single goal, and the tracking that adopts is when being multi-target method, perhaps working as moving target is single goal, and the tracking that adopts is when being multi-target method, and the error range between the measured value that obtains to the predicted value of the next position of moving target with by method for tracking target is usually all larger.

Estimated error can be weighed by dynamic switching reference value dsr.

In one embodiment, can calculate as follows dynamic switching reference value dsr:

At first, according to from the target status information of target following node, utilize the state equation of target and measurement equation to obtain predicted value and measured value to the next position of moving target by the information fusion node.Wherein the state equation of target and measurement equation are respectively

x(k+1)＝Φ(k+1，k)·X(k)+G(k)W(k)(4)

Y(k)＝H(k)·X(k)+V(k)(5)

X in the formula (k) ∈ R ^{N * 1}Be dbjective state vector, Y (k) ∈ R ^{M * 1}For measuring vector, W (k) ∈ R ^{P * 1}And V (k) ∈ R ^{M * 1}Be respectively state-noise and measurement noise, and be mutual incoherent white Gaussian noise sequence vector, its covariance matrix is respectively Q (k) and R (k), Φ (k+1, k) ∈ R ^{N * n}, G (k) ∈ R ^{N * p}And H (k) ∈ R ^{M * n}Be respectively state-transition matrix, input matrix and observing matrix.

Then, to the predicted value of next position and the estimating of the error between the measured value of moving target.In Kalman filtering method, as long as given initial value

And P (0/0), according to k measuring value Y (k) constantly, just can recursion calculate k state estimation constantly

And covariance matrix P (k/k).The fundamental equation of Kalman filtering is

$\hat{X}(k/k)=\hat{X}(k/k-1)+K\left(k\right)[Y\left(k\right)-H\left(k\right)X(k/k-1)]---\left(9\right)$

Wherein $\hat{X}(k/k-1)=\mathrm{\Φ}(k+1,k)\hat{X}(k-1/k-1)---\left(10\right)$

K(k)＝P(k/k-1)H ^T(k)[H(k)P(k/k-1)H ^T(k)+R(k)] ^-1 (11)

P(k/k-1)＝Φ(k+1，k)P(k-1/k-1)Φ(k+1，k)+G(k-1)Q(K-1)G ^T(k-1)(12)

P(k/k)＝[I-K(k)H(k)]P(k/k-1)(13)

Residual equation is $d\left(k\right)=Y\left(k\right)-H\left(k\right)\hat{X}(k/k-1)$

The d (k) that calculates by above-mentioned formula=(ξ ₁, ξ ₂..., ξ _n) ^T, representative is to the predicted value of the next position of moving target and the estimation of the error between the measured value.

Therefore, can utilize following formula to calculate dynamic switching reference value dsr tentatively to judge whether switching tracking:

$\mathrm{dsr}=\sqrt{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ξ}}_i}$ I=1 wherein, 2 ... n

In yet another embodiment, also can calculate dynamic switching reference value dsr by following method:

Can get target state equation X (t+1)=Φ X (t)+G ω (t) by the information fusion node according to the target status information from the target following node

Wherein $\mathrm{\Φ}=\left(\begin{array}{cc}1& T\\ 0& 1\end{array}\right),$ $G=\left(\begin{array}{c}{T}^2\\ 2\\ T\end{array}\right),$ T represents t constantly, and t+1 expresses the constantly next of t.State-noise ω (t) is the Gaussian sequence of zero-mean in the formula, its variance q=E[ω ²(t)].Measurement equation is y (t)=HX (t)+v (t)

H=[10 wherein], v (t) is that average is zero, variance r=E[v (t) v (t)] Gauss's white noise amount sound.The alpha-beta filtering equations is

$\hat{X}(t/t)=\hat{X}(t/t-1)+K[y\left(t\right)-H\hat{X}(t/t-1)]$

$\hat{X}(t/t-1)=\mathrm{\Φ}\hat{X}[(t-1)/(t-1)]$

$K=\left[\begin{array}{c}\mathrm{\α}\\ \mathrm{\β}/T\end{array}\right]$

Order $y\left(t\right)-H\left(t\right)\hat{X}(t/t-1)={({\mathrm{\χ}}_1,{\mathrm{\χ}}_2,{\mathrm{\χ}}_3,...,{\mathrm{\χ}}_n)}^T,$

Then can utilize following formula to calculate dynamic switching reference value dsr tentatively to judge whether switching tracking:

$\mathrm{dsr}=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\left|{\mathrm{\χ}}_j\right|$ J=1 wherein, 2 ... n

Step 3, judge that whether resulting dynamic switching reference value dsr is greater than predetermined threshold value ω, if greater than predetermined threshold value ω, then send message to cooperative nodes, determine current target number with the indication cooperative nodes based on the image via the collection of imageing sensor node.If the threshold value ω less than predetermined then can turn back to step 1).

Wherein, this threshold value ω can obtain by priori, for example establishes P (k/k)=(λ ₁, λ ₂..., λ _n) ^T, its two norm is so

Make ω=3 Γ.

In yet another embodiment, in order to quantize more accurately this judged result, and the convenient purpose that combines with the judged result of cooperative nodes, in step 3) can also calculate the credibility that based target is followed the tracks of this judgement that the information of node does, also can be called based target and follow the tracks of the confidence values that the information of node is judged the dynamic switching time of method for tracking target, for example be set as φ ₁, φ wherein ₁Computing formula be

0＜φ in the formula ₁＜1.

After step 4, cooperative nodes are received this message, judge current target number according to the image of current scene, and it is sent to the information fusion node as judged result, in order to judge comprehensively by the information fusion node whether target switches.

In one embodiment, cooperative nodes as follows, based on the moving target number of judging current scene via the image of the current scene of described imageing sensor node collection (also can referred to as current target number or target number).

At first, at cooperative nodes, read background image B _lWith present image I _l, the current goal that they gather from imageing sensor, wherein (x, y) represents pixel, l represents l width of cloth image.Obtain the absolute grayscale image of two two field pictures

d _l(x，y)＝|I _l(x，y)-B _l(x，y)|(16)

Secondly, setting threshold TH ₁, with the error image binaryzation, extract moving target

Then, use Mathematical Morphology Method to frame difference image b _lCarry out filtering and process, and then it is carried out the connectivity of region analysis, when the area S1 in the zone of a certain connection greater than a certain given threshold value σ 1, then become the detection target, and think that this zone is exactly the regional extent of target.

At last, calculate the number of such connected region, just can determine the number of target, be designated as h.If the number of target is 1 to be exactly single goal; If the number of target more than or equal to 2, is then thought multiple target.

In yet another embodiment, also can come by the following method to determine based on the image of the current scene of imageing sensor node collection the moving target number of current scene:

At first, extract two continuous two field picture IO by cooperative nodes _gAnd IO _G-1, wherein (X, Y) represents pixel, g represents g width of cloth image.Obtain the absolute difference gray level image D of two two field pictures _g

D _g(X，Y)＝|IO _g(X，Y)-IO _g-1(X，Y)|

Secondly, setting threshold TH ₂, with the error image binaryzation.When the difference of a certain pixel during greater than the threshold value set, think that then this pixel is object pixel, otherwise then think background pixel.

IB wherein _gBe bianry image, TH ₂Be the threshold value that binaryzation is set, (X, Y) represents pixel, and g represents g width of cloth image.

At last, if the area S2 in the zone that object pixel consists of then becomes the detection target greater than a certain given threshold value σ 2, calculate the number in the zone of satisfying this condition, if the number in such zone is 1 to be exactly single goal; If the number in zone is then thought multiple target more than or equal to 2 like this.

In yet another embodiment, in order to quantize more accurately this judged result, and the convenient purpose that combines with the judged result of cooperative nodes, in step 4) can also calculate the credibility of this judgement of being made by cooperative nodes, also can be called the confidence values that cooperative nodes is judged the dynamic switching time of method for tracking target, be assumed to be φ ₂, its computing formula is

0＜φ wherein ₂＜1, u=1,2 ... p, p represent the target number that current time detects, and q represents original target number, S _uThe area that represents the pixel region of u target, σ are the threshold values that is obtained by priori.

Cooperative nodes is based on after judging current target number or calculating the confidence values of this judgement via the image of imageing sensor node collection, its judged result is sent to the information fusion node, described judged result can be the current target number of being determined by cooperative nodes, or the confidence values that is calculated by cooperative nodes.

Step 5, by the information fusion node after the judged result of receiving from cooperative nodes, the judged result of the information of comprehensive based target tracking node and the judged result of cooperative nodes are made final judgement, namely judge the current method for tracking target that whether needs to switch, follow the tracks of node with indicating target and switch to corresponding method for tracking target.

The information fusion node is after the judged result of receiving from cooperative nodes, if dynamically switching reference value dsr is single goal greater than predetermined threshold value and current target number, then determine the current method for tracking target that needs to switch, follow the tracks of node with indicating target and switch to the monotrack method.If dynamically switching reference value dsr is multiple target greater than predetermined threshold value and current target number, then determine the current method for tracking target that needs to switch, follow the tracks of node with indicating target and switch to multi-object tracking method.

In yet another embodiment, for the comprehensive result who judges of quantitative information aggregators more accurately, in step 5) also can calculate this comprehensive confidence values of judging and determine whether and need to switch.Suppose that the comprehensive confidence values of judging is δ, its computing formula is as follows:

δ＝φ ₁+φ ₂

When the value of δ during greater than certain threshold value ψ, switch to corresponding target tracking algorism and carry out target following.For example, single goal when supposing beginning, and use the monotrack method and follow the tracks of, when the value of the comprehensive confidence values δ that obtains by cheap sensor node and imageing sensor node sometime during greater than predetermined threshold value ψ, just think present and occured from single goal to multiobject switching, vice versa.Wherein threshold value ψ pass through that priori obtains, general according to different scene spans between 1.2～1.8.

Step 6 judges whether to continue to follow the tracks of, and if so, then enters step 1, otherwise finishes.

In another example of the present invention, also provide a kind of system of dynamic switching method for tracking target.This system comprises target following node, cooperative nodes, information fusion node.Wherein the target following node sends to the information fusion node with the information that gathers, the confidence values that switching time is judged of the information that the information fusion node gathers based on the target following node and the information calculations that gathers based on cooperative nodes judges comprehensively whether the dynamic switching time of method for tracking target arrives.

Can find out by above-mentioned specific embodiment, the invention provides in Sensor Network target following scene, change the method for the dynamic switching of carrying out target tracking algorism along with target numbers.The Heterogeneous Sensor network that utilization is made of multiple sensors is sought the opportunity that target dynamic is switched.The method comprehensively judges whether to carry out the switching of method for tracking target in conjunction with the information from target following node and cooperative nodes.Use cheap sensor node to serve as the target following node, because it uses simply, computational speed is fast, can carry out real-time tracking to target.Simultaneously used again a small amount of imageing sensor node to serve as cooperative nodes, because data message is abundanter, it is more accurate to detect, but computational speed is slow, so can use imageing sensor in needs, to carry out IMAQ, with the detection destination number, and only carry out target following with cheap sensor node.Like this, take full advantage of the advantage of cheap sensor node and imageing sensor node, make the two collaborative work, the opportunity that can find more accurate, faster target to switch.

Although the present invention is described by preferred embodiment, yet the present invention is not limited to embodiment as described herein, also comprises without departing from the present invention various changes and the variation done.

Claims (10)

1. method that method for tracking target is dynamically switched that is used for sensor network, described method comprises:

Step 1) predicted value of the next position of moving target and the error between the actual measured value are estimated based on coming from the target status information of target following node by the information fusion node, dynamically to be switched reference value dsr;

Step 2) as resulting dynamic switching reference value dsr during greater than predetermined threshold value ω, sends message to cooperative nodes, determine current target number with the indication cooperative nodes;

Step 3) by cooperative nodes after the message of receiving from the information fusion node, judge current target number according to the image of current scene, and it sent to the information fusion node as judged result;

Step 4) by the information fusion node according to dynamically switching reference value and from the judged result of cooperative nodes, judging the current method for tracking target that whether needs to switch.

2. method according to claim 1, wherein, described cooperative nodes is imageing sensor.

3. method according to claim 1 is wherein, in step 1) described dynamic switching reference value dsr is:

ξ wherein _iBe i the component of vectorial d (k), d (k)=(ξ ₁, ξ ₂..., ξ _n) ^TThat the k moment is to the predicted value of the next position of moving target and the estimation of the error between the measured value.

4. method according to claim 3 is in step 2) also comprise if dsr, calculates based target greater than predetermined threshold value ω and follow the tracks of the confidence values φ that the information of node is judged the dynamic switching time of method for tracking target ₁Step, φ wherein ₁Computing formula be

0＜φ wherein ₁＜1.

5. method according to claim 4 is in step 3) also comprise and calculate the confidence values φ that cooperative nodes is judged the dynamic switching time of method for tracking target ₂, and send it to the step of information fusion node, wherein φ ₂Computing formula is

0＜φ wherein ₂＜1, u=1,2 ... p, p represent the target number that current time detects, and q represents original target number, S _uThe area that represents the pixel region of u target, σ are the threshold values of being scheduled to.

6. method according to claim 5 is in step 4) calculate confidence values δ by the information fusion node with following formula:

δ＝φ ₁+φ ₂

When the value of δ during greater than threshold value ψ, indicating target is followed the tracks of node and is switched to corresponding target tracking algorism and carry out target following.

7. method according to claim 6, wherein threshold value ψ according to different scene spans between 1.2～1.8.

8. method according to claim 1, wherein, in step 4) also comprise: as dynamic switching reference value dsr during greater than predetermined threshold value ω, if be single target from the current target number of cooperative nodes, then indicating target is followed the tracks of node and is switched to the monotrack algorithm; If the current target number from cooperative nodes is a plurality of targets, then indicating target is followed the tracks of node and is switched to the multiple target tracking algorithm.

9. system that method for tracking target is dynamically switched that is used for sensor network, described system comprises:

The target following node is used for the pursuit movement target, and sends target status information to the information fusion node;

The information fusion node is used for based on coming from the target status information of target following node to the estimating of error between the predicted value of the next position of moving target and the actual measured value, dynamically to be switched reference value dsr; And as resulting dynamic switching reference value dsr during greater than predetermined threshold value ω, send message to cooperative nodes, determine current target number with the indication cooperative nodes; And be used for according to dynamically switching reference value and from the judged result of cooperative nodes, judging the current method for tracking target that whether needs to switch;

Cooperative nodes is used for judging current target number according to the image of current scene, and it being sent to the information fusion node as judged result after the message of receiving from the information fusion node.

10. system according to claim 9, wherein, described cooperative nodes is imageing sensor.

Images