CN104735721B - Method based on compressed sensing technology image signal transmission in multi-hop wireless network - Google Patents

Abstract

The present invention provides a kind of method based on compressed sensing technology image signal transmission in multi-hop wireless network, including step：Original signal is packaged into the description packet that several information content are suitable, importance is identical by the first step, encoder using compressed sensing mode；Second step calculates the transmission benefit of every transmission link；Third walks, and is determined to need the link connection established according to the transmission benefit of transmission link；4th step establishes transmission link by scheduling scheme and starts transmission description packet.The present invention has considered coding and decoding and routing scheduling problem in picture signal transmission, and with very strong realizability, it can be widely deployed in wireless network transmissions till now, and then improve network capacity, be effectively improved picture signal transmission real-time and laser propagation effect.

Description

Method for image signal transmission based on compressed sensing technology in multi-hop wireless network

technical field

The present invention relates to a method in the technical field of wireless communication, in particular to an algorithm for encoding and packaging image data and routing and scheduling data packets for the purpose of optimizing network output, and specifically relates to compressive sensing technology based on multi-hop wireless networks A method of transmitting image signals.

Background technique

With the development of wireless network technology and the popularization of mobile devices, people's demand for mobile data services is increasing day by day. Therefore, big data and real-time mobile data transmission are attracting more and more attention. This topic has been extensively studied not only academically, but also in the field of industrial applications. Effectively using existing resources to provide users with mobile data services with better real-time performance and larger capacity is an important issue that needs to be solved urgently.

On the one hand, as the amount of data increases and the network load increases, the problem of routing and scheduling data packets will become more and more complicated. Compared with text signals, image signals (such as pictures, videos, etc.) have larger data volumes and higher real-time requirements, and image signals are just an indispensable part of improving user experience. The transmission requirements for transmitting image signals will increase will increase day by day. On the other hand, most of the previous inventions only considered the problem of encoding at the source end and decoding at the receiving end, or only considered the problem of routing and scheduling data packets in the network. A set of mutually adaptive algorithms for encoding, decoding and routing scheduling optimizes the transmission performance of image signals during encoding and transmission, which can improve network transmission effects and network transmission capabilities.

Contents of the invention

Aiming at the defects in the prior art, the purpose of the present invention is to provide a method to solve the problem of image signal transmission in a multi-hop wireless network. The invention comprehensively considers the problems of coding, decoding and routing scheduling in image signal transmission, has strong realizability, and can be widely used in current wireless network transmission. The invention can help us improve the network transmission capacity, and effectively improve the real-time and transmission effect of image signal transmission.

According to the method for transmitting an image signal based on compressed sensing technology in a multi-hop wireless network provided by the present invention, the method includes the following steps:

In the first step, the encoder uses compressed sensing to package the original signal into several description packages with the same amount of information and the same importance;

The second step is to calculate the transmission benefit of each transmission link;

The third step is to determine the link connection to be established according to the transmission benefit of the transmission link;

The fourth step is to establish the transmission link according to the scheduling scheme and start transmitting the description package.

Preferably, the step 1 is specifically:

First, the encoder performs M (M<N) sampling on the original image signal F of size N'×N' to obtain the sampled value y _m , where N=N'×N', F=[F _i,j ](i,j≤N'), F _i,j is the value on the corresponding point of the original image; the essence of sampling is to measure the original image signal F=[F _i,j ] in the sampling matrix φ' _m (m=1, 2,...,M) projection on

y _m = F, φ' _m

Among them, m=1,2,...,M represents the sampling times, the original image signal F=[F _i,j ] and the sampling matrix The size of each is N'×N';

Stack the two-dimensional image signal F=[F _i,j ] into one-dimensional column vector x=[x ₁ ,…x _N ],(n=1,2,…,N), where N=N '×N'; and the two-dimensional sampling matrix φ' _m is transposed by column and stacked into a one-dimensional row vector φ _m , then we have

y _m = F, φ' _m = φ _m x;

y=Φx

Among them, the sampling matrix Φ is a random matrix that obeys the same approximate distribution (i.i.d.) of ±1;

Then, the encoder packs the information of the sampling value y _m and its corresponding sampling matrix φ' _m into a description package;

The information of the sampling matrix Φ is shared between the sending end and the receiving end, that is, the sampling matrix Φ is pre-stored in the decoder of the sending end and the decoder of the receiving end, and the encoder only needs to indicate the sampling matrix in the description package when packaging the description package The number of samples m corresponding to φ' _m is sufficient.

Preferably, the second step is specifically:

(1) Calculate the transmission demand factor of the link

Assuming a certain time t, for a link l including sending node N _t and receiving node N _r , the set of description packets in the cache of sending node N _t is The set of description packets in the cache of the receiving node N _r is Assuming that each node can only forward the same received description packet at most once; then, there will be two different description packets in the node cache: one is the description packet received by the node and has been forwarded, called gray package; the other is the description package received by the node but not yet forwarded, which is called a white package; note that the collections of the gray package and the white package in the cache of the sending node N _t are respectively then there is

If link l is established, the packets available for transmission are those belonging to the set not part of the set The set of description packets; the set _UP of data packets available for transmission on the link l is

Considering the transmission capacity of the link, the number p of data packets that link l can transmit in each transmission time slot is

p=min(| _UP |,q(l))

Among them, q(l)=q(N _t , N _r ) represents the number of description packets that link l can transmit in a transmission time slot, and | _UP | represents the number of description packets contained in the set _UP ;

If link l is established, the reduction ΔRMS of the root mean square error of the restored image at the receiving node is

Wherein, the function f _RMS ( ) is the relationship between the root mean square error RMS of the compressed sensing restored image and the number of sampling values used for restoration;

Define the transmission demand factor TDF(l) of link l at this moment t as

The transmission demand factor of the link measures the importance of the data packets that the link will transmit;

(2) Calculate the transmission supply factor of the link;

(3) Calculate the transmission revenue and transmission cost of the link

Define the transmission revenue Revenue(l) of link l as the product of link transmission demand factor Revenue(l) and transmission supply factor TSF(l);

Revenue(l)=TDF(l)·TSF(l)

Define the transmission cost Cost(l) of link l as the transmission revenue of the link in the interference set of l, namely

Among them, the interference set S _int (l) of link l is the set of links that cannot work normally due to the interference brought by link l, namely

S _int (l)＝{l _int |l _int ≠l,Dis(N _t (l _int ),N _r (l))＜r,Dis(N _r (l _int ),N _t (l))＜r }

Among them, N _t (l) is the sending node of link l, N _r (l) is the receiving node of link l, Dis(N _{a ,} N _b ) is _the geographical distance or Channel quality, r is the interference radius of the node or the threshold of channel quality, l _int represents the link that will cause interference to link l;

(4) Obtain the transmission benefit of the link

The transmission benefit Utility(l) of link l is defined as

Preferably, the calculation of the transmission supply factor of the link is specifically:

First calculate the node weight of each relay node, specifically:

(a) The node performance value Q of each relay node is initialized to 0, the node weight value W is initialized to 0, and the hop number H from the node to the receiving node is initialized to positive infinity; the Q of the receiving node is positive infinity, and W is Positive infinity, H is 0;

(b) Traverse all nodes N _i in turn except the receiving node, if there is a node N _j directly connected to node N _i that satisfies

but

Q(N _i )←min(q(N _i ,N _j ),Q(N _j ))

H(N _i )←1+H(N _j )

That is, when N _i uses N _j as the next hop for transmission, and node N _i obtains a better node weight value, then update node N _i 's node performance value Q, node weight value W and the number of hops from the node to the receiving end node H; among them, q(l)=q(N _i , N _j ) represents the number of description packets that can be transmitted in a transmission time slot for a link l with N _i as the sending node and N _j as the receiving node, W( N _i ) represents the weight value of node N _i , Q(N _i ) represents the performance value of node N _i , H(N _i ) represents the number of hops from node N _i to the receiving end node;

(c) Repeat (b) until the node performance value Q of all nodes, the node weight value W and the hop number H from the node to the receiving node do not change in one traversal; get the node weight value W of each node;

The node weight measures the ability of each relay node to transmit data to the receiving end;

For a link l with N _r receiving nodes, the transmission supply factor TSF(l) of link l is defined as

TSF(l)=W(N _r )

Wherein, W(N _r ) represents the weight value of node N _r .

Preferably, the third step is specifically:

(1) At the beginning, links with positive transmission benefits are divided into links that may be established and added to the set of links to be established U _EstLink ;

(2) Find the link l _max with the largest transmission benefit, and then remove other links S _int (l _max ) that interfere with the link l _max from U _EstLink ; and reset the transmission benefit of the link l _max to zero, that is Utility(l _max )=0;

(3) Repeat (2) until the links in U _EstLink do not interfere with each other, that is Then U _EstLink is a set of links to be established in the time slot scheduling scheme.

Preferably, in the fourth step, the receiving end decodes the received description packet according to the actual situation.

min TV(F)

sty _r ＝Φ _r V _F

Wherein, the column vector V _F is formed by stacking the two-dimensional image signal F in columns, and TV(F) represents the total variation of the two-dimensional original image F corresponding to V _F .

Compared with the prior art, the present invention has the following beneficial effects:

(1) The present invention utilizes compressed sensing restoration technology, and the receiving end can decode the received description packets at any time without waiting for some missing description packets or requesting the sending end to retransmit, reducing the delay caused by waiting or retransmission operations. time delay.

(2) The present invention has strong realizability and pertinence for transmitting image signals with strong real-time performance and big data.

(3) In the routing scheduling algorithm of the present invention, the link with high transmission capacity will occupy the channel first. In this way, on the premise of avoiding channel conflicts, the entire network can have a higher output.

(4) The present invention comprehensively considers the problems of encoding, decoding and routing scheduling in image signal transmission. In this transmission method, the encoding and decoding of the image signal and the routing and scheduling of the data packet are adapted to each other, and the encoding and transmission cooperate with each other to optimize the transmission performance of the image signal, which can better improve the network transmission effect and improve the network transmission capacity.

(5) The present invention can ensure good data real-time performance and efficient routing scheduling, and when the link is established and transmitted, there will be no channel conflicts, and relatively good network output performance can be obtained at the same time.

Description of drawings

Other characteristics, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

Figure 1 is a schematic diagram of the comparison effect of the curves of the root mean square error (RMS) of the restoration images of the network scene 1 and the network scene 2 as a function of time.

FIG. 2 is a schematic diagram of the comparison effect of the curves of the number of data packets in the cache of the receiving end node in the network scenario 1 and the network scenario 2 as a function of time.

Detailed ways

The present invention will be described in detail below in conjunction with specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

The present invention provides a method for transmitting image signals based on compressed sensing technology in a multi-hop wireless network. The transmission method of the present invention includes the links of data packaging, data transmission, and data restoration that match each other, and can efficiently complete the image signal transmission in a wireless network. The problem of transmitting real-time and big data image signals in the medium. The steps of the present invention are as follows: first, the sending end packs the image data according to compressed sensing, then each node first calculates its own node weight, and then according to the state of the data packet in the cache of the nodes at both ends of the link and the receiving node in the link The node weight of the link is calculated to calculate the transmission benefit of the link, and the routing scheduling is carried out in units of links according to the link transmission benefit.

The network system model discussed in the present invention is a unicast network with a sending end, a corresponding receiving end, and multiple relay transmission nodes. Since the sending end and the receiving end are outside the communication range, the signal from the sending end needs to be forwarded by the relay node to reach the receiving end.

The present invention comprises the following steps:

In the first step, the encoder is responsible for packaging the original signal into several description packages with the same amount of information and the same importance by using compressed sensing technology.

First, the encoder will sample the original image signal F of size N'×N' for M (M<N) times to obtain the sampled value y _m , where N=N'×N', F=[F _{i ,j} ](i,j≤N'), F _i,j is the value on the corresponding point of the original image. The essence of sampling is to measure the projection of the original image signal F=[F _i,j ] on the sampling matrix φ' _m (m=1,2,...,M)

y _m = F, φ' _m

Among them, m=1,2,...,M represents the sampling times, the original image signal F=[F _i,j ] and the sampling matrix The size of each is N'×N'.

Stack the two-dimensional image signal F=[F _i,j ] into one-dimensional column vector x=[x ₁ ,…x _N ],(n=1,2,…,N), where N=N '×N'. And the two-dimensional sampling matrix φ' _m is transposed by column and stacked into a one-dimensional row vector φ _m , then we have

y _m =F, φ' _m =φ _m x.

y=Φx

Among them, the sampling matrix Φ is a random matrix that obeys ±1 equal probability distribution (i.i.d.).

Then, the encoder packs the information of the sampling value y _m and its corresponding sampling matrix φ' _m into a description package.

The information of the sampling matrix Φ is shared by the sending end and the receiving end, that is, the sampling matrix Φ is pre-stored in the decoder of the sending end and the decoder of the receiving end. Therefore, when the encoder packs the description package, it does not need to add the specific content of the sampling matrix to the description package, but only needs to indicate the sampling number m corresponding to the sampling matrix φ' _m in the description package. Therefore, the description packet is the data packet that contains the sample value and the number of samples.

In the second step, for each transmission link, calculate its transmission benefit. The specific steps of the algorithm are as follows:

(1) Calculate the transmission demand factor of the link.

Assuming a certain time t, for a link l including sending node N _t and receiving node N _r , the set of description packets in the cache of sending node N _t is The set of description packets in the cache of the receiving node N _r is Assume that each node can only forward the same received description packet at most once. Then, there will be two different description packets in the node cache: one is the description packet received by the node and has been forwarded, called the gray packet; the other is the description packet received by the node but not yet forwarded , known as the white bag. Note that the sets of gray packets and white packets in the cache of sending node N _t are then there is

If link l is established, the data packets available for transmission should not exist in the cache of the receiving node, and the white packets in the sending node; that is, those belonging to the set not part of the collection A collection of description packages. The set _UP of data packets available for transmission on link l is

Considering the transmission capacity of the link, the number p of data packets that link l can transmit in each transmission time slot is

p=min(| _UP |,q(l))

Among them, q(l)=q(N _t , N _r ) represents the number of description packets that link l can transmit in a transmission time slot, and | _UP | represents the number of description packets contained in the set _UP .

If link l is established, the reduction ΔRMS of the root mean square error of the restored image at the receiving node is

Among them, the function f _RMS (·) is the relationship between the root mean square error RMS of compressed sensing restored image and the number of sampling values used for restoration.

Define the transmission demand factor TDF(l) of link l at this moment t as

The transmission demand factor of a link measures the importance of the packets about to be transmitted by the link.

(2) Calculate the transmission supply factor of the link.

First calculate the node weight of each relay node, the calculation algorithm is

(a) The node performance value Q of each relay node is initialized to 0, the node weight value W is initialized to 0, and the hop number H from the node to the receiving node is initialized to positive infinity. The Q of the receiving end node is positive infinity, W is positive infinity, and H is 0.

(b) Traverse all nodes N _i in turn except the receiving node, if there is a node N _j directly connected to node N _i that satisfies

but

Q(N _i )←min(q(N _i ,N _j ),Q(N _j ))

H(N _i )←1+H(N _j )

That is, when N _i uses N _j as the next hop for transmission, and node N _i obtains a better node weight value, then update node N _i 's node performance value Q, node weight value W and the number of hops from the node to the receiving end node H. Among them, q(l)=q(N _i , N _j ) represents the number of description packets that can be transmitted in a transmission time slot for a link l with N _i as the sending node and N _j as the receiving node, W(N _i ) represents the weight value of node N _i , Q(N _i ) represents the performance value of node N _i , and H(N _i ) represents the number of hops from node N _i to the receiving end node.

(c) Repeat (b) until the node performance value Q of all nodes, the node weight value W and the hop number H from the node to the receiving node do not change in one traversal. Get the node weight value W of each node.

The node weight measures the ability of each relay node to transmit data to the receiver.

For a link l with N _r receiving nodes, its transmission supply factor TSF(l) is defined as

TSF(l)=W(N _r )

Wherein, W(N _r ) represents the weight value of node N _r .

(3) Calculate the transmission revenue and transmission cost of the link.

The transmission revenue Revenue(l) of link l is defined as the product of link transmission demand factor Revenue(l) and transmission supply factor TSF(l).

Revenue(l)=TDF(l)·TSF(l)

Define the transmission cost Cost(l) of link l as the transmission revenue of the link in the interference set of l, namely

The interference set S _int (l) of link l is the set of links that cannot work normally due to the interference brought by link l, namely

S _int (l)＝{l _int |l _int ≠l,Dis(N _t (l _int ),N _r (l))＜r,Dis(N _r (l _int ),N _t (l))＜r }

Among them, N _t (l) is the sending node of link l, N _r (l) is the receiving node of link l, Dis(N _{a ,} N _b ) is _the geographical distance or Channel quality, r is the interference radius of the node or the threshold of channel quality, l _int indicates the link that will cause interference to link l.

(4) Get the transmission benefit of the link.

The transmission benefit Utility(l) of link l is defined as

The third step is to determine the link connection to be established according to the transmission efficiency of the link. The specific algorithm is as follows:

(1) The set of links to be established is U _EstLink . Initially, all links with positive transfer benefits are divided into

Links that may be established are added to the set U _EstLink .

(2) Find out the link l _max with the largest transmission benefit, and then remove other links S _int (l _max ) that interfere with the link l _max from U _EstLink . And the transmission benefit of the link l _max is reset to zero, that is, Utility(l _max )=0. Since the link interfering with l _max has been removed at this stage, l _max will not be affected in subsequent iterations.

(3) Repeat (2) until the links in U _EstLink do not interfere with each other, that is Then U _EstLink is a set of links to be established in the time slot scheduling scheme.

After the above algorithm traverses all the links, all possible interferences are shielded, and the system reaches a stable transmission state. This stable transmission state means that within each interference range, the channel is occupied by a link with a locally optimal transmission utility value under the premise of no interference.

In the fourth step, the transmission link is established according to the scheduling scheme and the description packet is started to be transmitted. The receiving end can decode the received description packet according to the actual situation.

min TV(F)

sty _r ＝Φ _r V _F

Wherein, the column vector V _F is formed by stacking the two-dimensional image signals F in columns.

In the experiment, the wireless network is distributed in a square area of 80*80, and the midpoints on the opposite sides of the square are a sending node and a receiving node respectively; in addition, 18 relay nodes are randomly distributed in the square area. Both the interference radius and the communication radius of wireless communication are r=30 meters, that is, two nodes within a radius of 30 meters can be affected by wireless signals sent by each other. The signal-to-noise ratio of a communication link between two nodes within the communication range satisfies a Gaussian distribution. At the same time, we know that within the interference radius of the transmitting node, other nodes except the receiving node cannot receive data; within the interference radius of the receiving node, other nodes except the transmitting node cannot send data. And any node can only receive data from one node at the same time, and cannot receive and send data at the same time. According to these relationships, we can establish the interference relationship matrix between links in this network scenario.

The specific implementation steps include the following steps:

The first step is to calculate the node weight value W of each relay node according to the network topology relationship, and set the node weight value W of the receiving end node to 100 (there is no special numerical requirement for this 100, but the node weight value of the receiving end node needs to be relatively The node weight value of other nodes is sufficiently large).

The second step is to calculate the transmission demand factor of each link, the transmission revenue, transmission cost and transmission benefit of each link according to the algorithm we introduced.

The third step is to determine the link connection that needs to be established according to the transmission efficiency of the link, establish a communication link, and update the status of the data packet in the cache of each node.

The fourth step is to record the following data:

(1) The number of data packets received by the receiving end node in each time slot.

(2) In each time slot, the mean square error value of the restoration graph of the signal restoration performed by the receiving node according to the received data packets relative to the original graph.

In two different network scenarios, by recording the above data and comparing the experimental results, we obtained Figure 1 and Figure 2. Among them, the content of the comparative experiment is to transmit the same image signal in the same wireless network using the transmission method based on the traditional discrete cosine transform.

In Figure 1, (a) is the restoration effect of network scene 1, and (b) is the restoration effect of network scene 2. In Figure 2, (a) is the restoration effect of network scene 1, and (b) is the restoration effect of network scene 2.

Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention.

Claims (5)

1. A method for transmitting image signals based on compressed sensing technology in a multi-hop wireless network, characterized in that, comprising the steps: In the first step, the encoder uses compressed sensing to package the original signal into several description packages with the same amount of information and the same importance; The second step is to calculate the transmission benefit of each transmission link; The third step is to determine the link connection to be established according to the transmission benefit of the transmission link; The fourth step is to establish the transmission link according to the scheduling scheme and start transmitting the description package; The third step is specifically: (1) At the beginning, links with positive transmission benefits are divided into links that may be established and added to the set U _EstLink of links to be established; (2) Find the link l _max with the largest transmission benefit, and then remove other links S _int (l _max ) that interfere with the link l _max from U _EstLink ; and reset the transmission benefit of the link l _max to zero, namely Utility(l _max )=0; (3) Repeat (2) until the links in U _EstLink do not interfere with each other, that is Then U _EstLink is the set of links to be established by the scheduling scheme; The second step is specifically: (1) Calculate the transmission demand factor of the link Assuming a certain time t, for a link l including sending node N _t and receiving node N _r , the set of description packets in the cache of sending node N _t is The set of description packets in the cache of the receiving node N _r is Assuming that each node can only forward the same received description packet at most once; then, there will be two different description packets in the node cache: one is the description packet received by the node and has been forwarded, called gray package; the other is the description package received by the node but not yet forwarded, which is called a white package; note that the collections of the gray package and the white package in the cache of the sending node N _t are respectively then there is If link l is established, the packets available for transmission are those belonging to the set not part of the collection The set of description packets; the set _UP of data packets available for transmission on the link l is Considering the transmission capacity of the link, the number p of data packets that link l can transmit in each transmission time slot is p=min(|U _P |,q(l)) Among them, q(l)=q(N _t , N _r ) represents the number of description packets that link l can transmit in a transmission time slot, and | _UP | represents the number of description packets contained in the set _UP ; If link l is established, the reduction ΔRMS of the root mean square error of the restored image at the receiving node is Wherein, the function f _RMS ( ) is the relationship between the root mean square error RMS of the compressed sensing restored image and the number of sampling values used for restoration; Define the transmission demand factor TDF(l) of link l at this moment t as The transmission demand factor of the link measures the importance of the data packets that the link will transmit; (2) Calculate the transmission supply factor of the link; (3) Calculate the transmission revenue and transmission cost of the link Define the transmission revenue Revenue(l) of link l as the product of link transmission demand factor TDF(l) and transmission supply factor TSF(l); Revenue(l)=TDF(l)·TSF(l) Define the transmission cost Cost(l) of link l as the transmission revenue of the link in the interference set of l, namely Among them, the interference set S _int (l) of link l is the set of links that cannot work normally due to the interference brought by link l, namely S _int (l)＝{l _int |l _int ＝l,Dis(N _t (l _int ),N _r (l))＜r,Dis(N _r (l _int ),N _t (l))＜r } Among them, N _t (l) is the sending node of link l, N _r (l) is the receiving node of link l, Dis(N _{a ,} N _b ) is _the geographical distance or Channel quality, r is the interference radius of the node or the threshold of channel quality, l _int represents the link that will cause interference to link l; (4) Get the transmission benefit of the link The transmission benefit Utility(l) of link l is defined as 2. The method for transmitting image signals based on compressed sensing technology in a multi-hop wireless network according to claim 1, wherein the first step is specifically: First, the encoder performs M (M<N) sampling on the original image signal F of size N'×N' to obtain the sampled value y _m , where N=N'×N', F=[F _i,j ](i,j≤N'), F _i,j is the value on the corresponding point of the original image; the essence of sampling is to measure the original image signal F=[F _i,j ] in the sampling matrix φ' _m (m=1, 2,...,M) projection on y _m ＝<F,φ' _m > Among them, m=1,2,...,M represents the sampling times, the original image signal F=[F _i,j ] and the sampling matrix The size of each is N'×N'; Stack the two-dimensional image signal F=[F _i,j ] into one-dimensional column vector x=[x ₁ ,…x _N ],(n=1,2,…,N), where N=N '×N'; and the two-dimensional sampling matrix φ' _m is transposed by column and stacked into a one-dimensional row vector φ _m , then we have y _m =<F,φ' _m >=φ _m x; Stack ym into a matrix y of size M×1, namely The inversion φ _m ^T of the sampling matrix φ _m is stacked by rows into a matrix Φ of size M×N, namely then there is y=Φx Among them, the sampling matrix Φ is a random matrix that obeys the same approximate distribution (i.i.d.) of ±1; Then, the encoder packs the information of the sampling value y _m and its corresponding sampling matrix φ' _m into a description package; The information of the sampling matrix Φ is shared by the sending end and the receiving end, that is, the sampling matrix Φ is pre-stored in the decoder of the sending end and the decoder of the receiving end, and the encoder only needs to indicate the sampling matrix in the description packet when packaging the description packet. The number of samples m corresponding to φ' _m is sufficient. 3. The method for transmitting image signals based on compressed sensing technology in a multi-hop wireless network according to claim 1, wherein the calculation of the transmission supply factor of the link is specifically: First calculate the node weight of each relay node, specifically: (a) The node performance value Q of each relay node is initialized to 0, the node weight value W is initialized to 0, and the hop number H from the node to the receiving node is initialized to positive infinity; the Q of the receiving node is positive infinity, and W is Positive infinity, H is 0; (b) Traverse all nodes N _i in turn except the receiving node, if there is a node N _j directly connected to node N _i that satisfies but Q(N _i )←min( _q (N _i ,N _j ),Q(N _j )) H(N _i )←1+H(N _j ) That is, when N _i uses N _j as the next hop for transmission, and node N _i obtains a better node weight value, then update node N _i 's node performance value Q, node weight value W and the number of hops from the node to the receiving end node H; Among them, q(l)=q(N _i , N _j ) represents the number of description packets that can be transmitted in a transmission time slot for a link l with N _i as the sending node and N _j as the receiving node, W( N _i ) represents the weight value of the node N _i , Q(N _i ) represents the performance value of the node N _i , and H(N _i ) represents the number of hops from the node N _i to the receiving end node; (c) Repeat (b) until the node performance value Q of all nodes, the node weight value W and the hop number H from the node to the receiving node no longer change in one traversal; obtain the node weight value W of each node; The node weight measures the ability of each relay node to transmit data to the receiving end; For a link l with N _r receiving nodes, the transmission supply factor TSF(l) of link l is defined as TSF(l)=W(N _r ) Wherein, W(N _r ) represents the weight value of node N _r . 4. The method for transmitting image signals based on compressed sensing technology in a multi-hop wireless network according to claim 1, characterized in that, in the fourth step, the receiving end carries out the description packet received according to the actual situation decoding. 5. The method for transmitting image signals based on compressed sensing technology in a multi-hop wireless network according to claim 4, characterized in that, in the fourth step, if at time t, the receiving end receives K corresponding samples Description packets with degrees {k ₁ ,k ₂ ,...,K}, where Let the sample value contained in the description package be Stacked by number of samples as The corresponding sampling matrix is Restore the original signal, that is, solve the optimization problem min TV(F) sty _r ＝Φ _r V _F Wherein, the column vector V _F is formed by stacking the two-dimensional image signal F in columns, and TV(F) represents the total variation of the two-dimensional original image F corresponding to V _F .