RFID target imaging method based on line segment subdivision and holographic positioning

Abstract

An RFID target imaging method based on line segment subdivision and holographic positioning comprises the following steps: collecting a tag phase under the condition that a target exists or not, and calculating an internal propagation distance of the target; acquiring direct path propagation distance and arrival angle data through the signal capture points and the label coordinates; establishing a phase calculation model, and obtaining a propagation distance starting point phase; obtaining the theoretical phase data of the sub-line segments by using a polar coordinate system subdivision model, and carrying out holographic positioning through the phase difference between the starting point and the sub-line segments; determining an end point according to the starting point and the propagation distance under the polar coordinates, and converting the end point into Cartesian coordinates; and optimizing the starting point and the ending point by adopting a chicken flock optimization algorithm. According to the invention, by mining the relationship between the propagation distance starting point and the sub-line segment, the problem of poor imaging effect of the arc-shaped target is improved, and the imaging precision and efficiency are improved.

Description

Technical Field

[0001] The invention belongs to the field of wireless communications and relates to an ultra-high frequency RFID target imaging method based on line segmentation and holographic positioning. Background Art

[0002] With the rapid development of science and technology, target imaging technology has become a core component of modern information acquisition and processing, and continues to receive high attention from academia and industry. Especially in complex application environments, such as medical diagnosis, industrial detection, security monitoring and other fields, related technologies and theories have become a research hotspot for scientific researchers. As a component of imaging technology, RFID target imaging technology combines the advantages of radio frequency identification and imaging technology, providing innovative solutions for object imaging. Its non-line-of-sight imaging capability can effectively overcome the system's demand for light brightness, while having good privacy protection capabilities; the good penetration ability of radio frequency signals increases the scope of application of the system; the low cost and low power consumption characteristics make RFID target imaging technology stand out from many technologies, and it can quickly and accurately locate and identify multiple targets in complex environments, effectively solving the problems of obstructed line of sight and much manual intervention in traditional methods, and greatly improving management efficiency and system automation level.

[0003] A typical RFID system generally consists of a reader, an antenna, and a data acquisition system built with several tags. In the existing target imaging method, the target contour candidates are only linear. When imaging circular or elliptical arc-shaped targets, the estimated image deviates greatly from the actual image. In addition, the existing target imaging method cannot capture enough spatial features of the target, resulting in problems in selecting the starting point of the propagation distance, and the estimated image does not match the actual image. In response to these problems, the present invention proposes an RFID target imaging method based on segmentation and holographic positioning. Summary of the invention

[0004] The purpose of the present invention is to propose an RFID target imaging method based on line segmentation and holographic positioning, so as to solve the problem that the imaging image of circular or elliptical arc-shaped targets in the existing imaging method has a large deviation from the actual image. The technical solution adopted by the present invention is to use ultra-high frequency RFID technology to obtain the tag phase in both the case of no target and the case of target and take the difference between the two, establish a distance starting point phase calculation model to obtain the phase of the propagation distance starting point; establish a polar coordinate system with the tags in two directions as poles, establish a direct path line segmentation model to narrow the search range of the propagation distance starting point holographic positioning, obtain the theoretical phase data of the sub-segment, use the phase value between the propagation distance starting point and the sub-segment for holographic positioning, and use the chicken swarm optimization algorithm to optimize the propagation distance starting point, and finally aggregate all target edge points into a hologram to achieve horizontal section image reconstruction, which specifically includes the following steps:

[0005] Step 1: Build a system through RFID readers, antennas, tags and controllable mobile robots to collect tag phases and obtain the direct path propagation distance and angle of arrival (AOA) data of the signal. Establish a two-dimensional Cartesian coordinate system in the detection area with a length of L and a width of W, and set the controllable mobile robot to start from the initial position (0, W) at a speed v and move along a given trajectory. Set the coordinates of the two tags to (L / 2, 0) and (0, W / 2) respectively, define the vertical direction of tag T1 as direction 1, and define the horizontal direction of tag T2 as direction 2. Collect the tag phases in both the case of no target and with target, and take the difference between the two to obtain two sets of phase differences △φ. x and ∆φ y The cumulative phase change method based on sliding window is used to detect the target contour boundary, and the phase data collected within the boundary is retained and recorded as △φ′ x and ∆φ′ y , the data volume is recorded as V x and V y , using the phase △φ′ x and ∆φ′ y Calculate the propagation distance PD of the signal inside the target x and PD y , the calculation formula is PD = (△φ + 2σπ) / (β tar -β air ), where β air and β tar They represent the phase constants of the signal in the air and in the target respectively. Since the phase change caused by the signal penetrating the target generally does not exceed 2π, σ is usually O. The signal capture point position AP can be obtained from the robot's moving speed and the timestamp when the signal is captured by the antenna. x and AP y , assuming and Respectively represent the coordinates of a signal capture point in direction 1 and direction 2. According to the formula Get the length DL of the direct path line segment of the signal in both directions x ,DL y , according to the formula Get AOA dataθ x ,θ y ;

[0006] Step 2: Obtain the theoretical phase of the starting point of the propagation distance. Take the mth propagation distance starting point in direction 1 as an example. Assume that the distance between the mth propagation distance starting point and its corresponding signal capture point is x. m , then the distance between it and the label is The length of the i-th direct path segment is recorded as i∈[1,V x ] and i≠m, the angle between the i-th direct path segment and the direct path segment where the m-th propagation distance starting point is located is recorded as i∈[1,V x ] and i≠m, using the formula The distance between the starting point of the mth propagation distance and the other i signal capture points can be obtained as Then E m Substituting into the expression φ m =(4πE m / λ)mod2π, the theoretical phase value φ of the starting point of the mth propagation distance under ideal conditions can be obtained. m , and so on, we can get the theoretical phase value matrix of all propagation distance starting points in two directions and

[0007] Step 3: Construct a direct path segment decomposition model in the polar coordinate system, calculate the theoretical phase value of each direct path segment sub-segment, take the label 1 direction as an example, take the label 1 position (L / 2, 0) as the polar coordinate reference system pole O1, take the positive direction of the Cartesian coordinate system abscissa as the polar axis of the polar coordinate system, and use the polar diameter r and polar angle represents the coordinates of each point in the imaging area in the polar coordinate system, then each point on the direct path segment O1A can be To indicate that, It is obtained by converting the AOA data of the direct path segment to r f is the segment step length. The idea of ​​segment segmentation is used to segment O1A into several sub-segments. The coordinates of the midpoint of the sub-segment are the coordinates of the sub-segment. In the same polar coordinate system, the position A of the signal capture point x Can be used To represent the coordinates, where r j The length Dt of each direct path segment x Assume that the coordinates of the midpoint Q of any sub-segment T of O1A in the polar coordinate system are The location of any signal capture point The coordinates of The distance between the two You can use the formula Calculated, among which, And j∈[1,V x ], Then use the formula 2π can get the signal from The phase value measured when the light is reflected back through the sub-line segment T By analogy, the theoretical phase value set of all direct path line segments and sub-line segments in two directions can be obtained: and

[0008] Step 4: Use the idea of ​​holographic positioning to complete the positioning of the starting point of the propagation distance. and Respectively and By comparison, we can obtain the probability that each sub-segment is the starting point of the propagation distance, and use the formula Calculate the pixel value for quantization, where T i,j Represents the jth sub-segment on the i-th direct path segment. For a direct path segment, find the sub-segment with the largest pixel value belonging to the direct path segment and the coordinates of the starting point of the propagation distance It is the coordinate of the midpoint of the sub-segment with the largest pixel value, and the same is true for other direct path segments;

[0009] Step 5: Construct an optimization model based on maximizing overlap to find the optimal starting point of the propagation distance. The coordinates of the end point of the propagation distance can be obtained based on the coordinates of the starting point of the propagation distance, the propagation distance, and the corresponding AOA data. The calculation formula for the coordinates of the end point of the propagation distance in the tag 1 direction is: Among them, i∈[1,V x ], The calculation formula for the coordinates of the end point of the propagation distance in the tag 2 direction is: Among them, i∈[1,V y ], Convert the polar coordinates of the starting and ending points of the propagation distance into Cartesian coordinates. The conversion formula for label 1 direction is: The formula for label 2 direction conversion is: The starting point and the end point of the propagation distance are connected in sequence in different directions to form two closed figures connected by points and lines. The closed figures are filled with colors to obtain two estimated images. After merging, the ratio of the overlapping area to the total area is compared to obtain the overlap of the two estimated images. The overlap of the estimated images in the two directions is maximized by adjusting the distance X between the starting point of the propagation distance and the signal capture point. The maximum number of iterations set is used as the termination condition to construct an optimization model for the starting point of the propagation distance. The swarm optimization algorithm is introduced to solve the problem, so as to realize the fast and accurate search of the starting point and the end point of the propagation distance.

[0010] Step 6: Aggregate all the optimal propagation distance starting points and ending points into a hologram to generate a composite hologram. Use the composite hologram reconstruction method based on the target prior information to present the horizontal section image of the target. According to the imaging scene deployment and data acquisition method, perform holographic positioning of the optimal propagation distance starting point and ending point in the plane grid area to obtain the pixel values ​​of the propagation distance starting point and ending point in the plane grid area. All the propagation distance starting points and ending points are collectively referred to as target edge points. Use the formula PX normalization =PX / PX max Normalize all pixel values ​​of a single target edge point, where PX represents the unnormalized pixel value matrix of a target edge point, PX max Indicates the largest pixel value in the pixel value matrix, performs color mapping on pixel values ​​of different sizes and fills the corresponding grid to form a hologram of a single target edge point, uses a weight coefficient ω ranging from 0 to 1 to suppress the pixel values ​​of the non-target edge point grid, establishes a set of target edge point holograms in two label directions, superimposes the holograms contained in a single direction, obtains a single-direction sub-hologram of the target horizontal section, and superimposes the sub-holograms in two directions to generate a composite hologram of the target horizontal section;

[0011] It should be noted that the basis for holographic positioning of the starting point and the end point of the propagation distance in step 6 is that the sub-segments on each direct path segment belong to different angular directions, and when the same segment step size is used to divide each direct path segment, the number of sub-segments in each direction is not exactly the same, so the target edge point hologram based on segment division cannot be superimposed. In addition, all target edge point holograms based on segment division have no intersection in space, and simply superimposing them cannot reconstruct the horizontal section image of the target. Therefore, all target edge points need to be holographically positioned to obtain their pixel values ​​in the plane grid area, and then generate a hologram based on plane division to complete image reconstruction. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a system framework diagram of the present invention;

[0013] Figure 2 It is a schematic diagram of the scene of the present invention;

[0014] Figure 3 It is a geometric model diagram of the phase calculation of the starting point of the propagation distance in the direction of tag 1 of the present invention;

[0015] Figure 4 It is a trigonometric relationship diagram of the cosine theorem of the present invention;

[0016] Figure 5 It is a geometric model diagram of the direct path segment decomposition in the polar coordinate system of the label 1 direction of the present invention;

[0017] Figure 6 is the target composite hologram of the present invention when the weight coefficient is ω=0.1;

[0018] Figure 7 It is a deviation curve diagram of two arc-shaped targets of the present invention;

[0019] Figure 8 It is the deviation curve diagram of two arc-shaped targets under three different noise environments of the present invention;

[0020] Fig. 9 It is a comparison chart of the deviation curves of the imaging target under different line segment step sizes. DETAILED DESCRIPTION

[0021] The purpose of the present invention is to propose an RFID target imaging method based on line segmentation and holographic positioning. The method takes into account that the edge point of the target to be positioned is located in the area between the tag and the antenna, establishes a direct path line segmentation model in a polar coordinate system, narrows the search range of holographic positioning from the grid area of ​​the entire plane to the direct path line segment, uses swarm optimization to search for the optimal propagation distance starting point with maximizing the overlap of estimated images in two directions as the optimization direction, and finally uses holographic positioning to obtain the pixel values ​​of the optimal propagation distance starting point and end point. After processing, a hologram of a single target edge point is generated, all target edge points are aggregated, and a composite hologram based on target prior information is used to present the target horizontal section image.

[0022] The following is combined with Figure 1 , Attachment Figure 2 , Attachment Figure 3 , Attachment Figure 4 , Attachment Figure 5 , Attachment Figure 6 , Attachment Figure 7 , Attachment Figure 8 , Attachment Fig. 9 The embodiments of the present invention are described in further detail.

[0023] The system framework of the present invention is shown in the attached Figure 1 As shown, we first build an ultra-high frequency RFID target imaging system. The system framework mainly includes: offline and online phase data collection and preprocessing, establishing a direct path segmentation model under polar coordinates, using the swarm optimization algorithm to find the optimal propagation distance starting point, and presenting the target horizontal section image based on the composite hologram of the target prior information. It is mainly composed of readers, antennas, tag arrays, background processors and other equipment.

[0024] According to the attached Figure 2After completing the construction of the system scenario, collect the tag phase and obtain the signal direct path propagation distance and arrival angle (AOA) data. Use RFID readers, antennas, tags and controllable mobile robots to build a system to collect tag phases. Establish a two-dimensional Cartesian coordinate system in the detection area with a length of L and a width of W. Set the controllable mobile robot to start from (0, W) at a speed v and move along a given trajectory. Set the coordinates of the two tags to (L / 2, 0) and (0, W / 2) respectively. Define the vertical direction of tag T1 as direction 1, and define the horizontal direction of tag T2 as direction 2. Collect the tag phases in both the no-target and target situations and take the difference between the two to obtain two sets of phase differences △φ x and ∆φ y , the cumulative phase change method based on sliding window is used to detect the target contour boundary, and the phase data collected within the boundary is retained and recorded as △φ′ x and ∆φ′ y , the data volume is recorded as V x and V y , using the phase △φ′ x and ∆φ′ y Calculate the propagation distance PD of the signal inside the target x and PD y , the calculation formula is PD = (△φ + 2σπ) / (β tar -β air ), where β air and β tar They represent the phase constants of the signal in the air and in the target respectively. Since the phase change caused by the signal penetrating the target generally does not exceed 2π, σ is usually 0. The signal capture point position AP can be obtained from the robot's moving speed and the timestamp when the signal is captured by the antenna. x and AP y , assuming and Respectively represent the coordinates of a signal capture point in direction 1 and direction 2. According to the formula Get the length DL of the direct path line segment of the signal in both directions x ,DL y , according to the formula Get AOA dataθ x ,θ y Next, obtain the theoretical phase of the propagation distance starting point. Take the mth propagation distance starting point in direction 1 as an example. Assume that the distance between the mth propagation distance starting point and its corresponding signal capture point is x. m , then the distance between it and the label is The length of the i-th direct path segment is recorded as i∈[1,V x] and i≠m, the angle between the i-th direct path segment and the direct path segment where the m-th propagation distance starting point is located is recorded as i∈[1,V x ] and i≠m, using the formula The distance between the starting point of the mth propagation distance and other signal capture points can be obtained Then E m Substituting into the expression φ m =(4πE m / λ)mod 2π, the theoretical phase value φ at the starting point of the mth propagation distance under ideal conditions can be obtained. m , and so on, we can get the theoretical phase value matrix of all propagation distance starting points in two directions and Attached Figure 3 The figure shows the geometric model for calculating the phase at the starting point of the propagation distance in the direction of tag 1. Figure 4 Shown is the cosine theorem for solving trigonometric relationships.

[0025] Construct a direct path segment decomposition model in the polar coordinate system, calculate the theoretical phase value of each direct path segment sub-segment, take the label 1 direction as an example, take the label 1 position (L / 2, 0) as the polar coordinate reference system pole O1, take the positive direction of the Cartesian coordinate system abscissa as the polar axis of the polar coordinate system, and use the polar diameter r and polar angle represents the coordinates of each point in the imaging area in the polar coordinate system, then each point on the direct path segment O1A can be To indicate that, It is obtained by converting the AOA data of the direct path segment to r f is the segment step length. The idea of ​​segment segmentation is used to segment O1A into several sub-segments. The coordinates of the midpoint of the sub-segment are the coordinates of the sub-segment. In the same polar coordinate system, the position A of the signal capture point x Can be used To represent the coordinates, where r j The length Dt of each direct path segment x Assume that the coordinates of the midpoint Q of any sub-segment T of O1A in the polar coordinate system are The location of any signal capture point The coordinates of The distance between the two You can use the formula Calculated, among which, And j∈[1,V x ], Then use the formula The signal can be obtained from The phase value measured when the light is reflected back through the sub-line segment T By analogy, the theoretical phase value set of all direct path line segments and sub-line segments in two directions can be obtained: and Then and Respectively and By comparison, we can obtain the probability that each sub-segment is the starting point of the propagation distance, and use the formula Calculate the pixel value for quantization, where T i，j Represents the jth sub-segment on the i-th direct path segment. For a direct path segment, find the sub-segment with the largest pixel value belonging to the direct path segment and the coordinates of the starting point of the propagation distance is the midpoint coordinate of the sub-segment with the largest pixel value, and the same is true for other direct path segments. Figure 5 Shown is the geometric model of the direct path segment decomposition in polar coordinates.

[0026] Construct an optimization model based on maximizing overlap to find the optimal starting point of the propagation distance. The coordinates of the end point of the propagation distance can be obtained according to the coordinates of the starting point of the propagation distance, the propagation distance and the corresponding AOA data. The calculation formula for the coordinates of the end point of the propagation distance in the tag 1 direction is: Among them, i∈[1,V x ], The calculation formula for the coordinates of the end point of the propagation distance in the tag 2 direction is: Among them, i∈[1,V y ], Convert the polar coordinates of the starting and ending points of the propagation distance into Cartesian coordinates. The conversion formula for label 1 direction is: The formula for label 2 direction conversion is: The obtained starting point and ending point of the propagation distance are connected in sequence in different directions to form two closed figures connected by points and lines. The closed figures are filled with colors to obtain two estimated images. After merging, the ratio of their overlapping area to the total area is compared to obtain the overlap of the two estimated images. The overlap of the estimated images in the two directions is maximized by adjusting the distance X between the starting point of the propagation distance and the signal capture point. The set maximum number of iterations is used as the termination condition, and an optimization model for the starting point of the propagation distance is constructed. The swarm optimization algorithm is introduced for solving the problem to achieve fast and accurate search for the starting point and ending point of the propagation distance. Finally, all the optimal starting points and ending points of the propagation distance are aggregated into a hologram to generate a composite hologram. The horizontal section image of the target is presented using the composite hologram reconstruction method based on the target prior information. According to the scene deployment and data acquisition method of the imaging, the optimal starting point and ending point of the propagation distance are holographically located in the plane grid area to obtain the pixel values ​​of the starting point and ending point of the propagation distance in the plane grid area. All the starting points and ending points of the propagation distance are collectively referred to as target edge points. The formula PX is used to calculate the target edge points. normalization =PX / PX max Normalize all pixel values ​​of a single target edge point, where PX represents the unnormalized pixel value matrix of a target edge point, PX max Indicates the maximum pixel value in the pixel value matrix, performs color mapping on pixel values ​​of different sizes and fills the corresponding grid to form a hologram of a single target edge point, uses a weight coefficient ω ranging from 0 to 1 to suppress the pixel value of the non-target edge point grid, establishes a set of target edge point holograms in two label directions, superimposes the holograms contained in a single direction, obtains a single-direction sub-hologram of the target horizontal section, superimposes the sub-holograms in two directions, and generates a composite hologram of the target horizontal section. Figure 6 Shown is the target composite hologram when the weight coefficient is ω=0.1.

[0027] The following is the specific implementation: Figure 2As shown in the figure, the reader antenna uses a linearly polarized microstrip antenna, and the tag uses a dipole antenna. The commercial RFID device Impinj R420 and the programmable robot iRobot are used to collect the tag phase data. Specifically, a two-dimensional Cartesian coordinate system is established in the detection area with a length of L and a width of W. The coordinates of the two tags are (L / 2, 0) and (0, W / 2) respectively, and the vertical direction of the tag T1 is defined as direction 1, and the horizontal direction of the tag T2 is defined as direction 2. The robot starts from (0, W) at a speed v and moves along a given trajectory. The robot's motion trajectory is to first move horizontally from (0, W) at a speed v, and then move vertically at the same speed V after moving to the position (L, W) and finally reach the position (L, 0). According to the above scenario, the controllable mobile robot is set to collect the tag phase in the two cases of no target and with target respectively and take the difference between the two. The cumulative phase change method based on the sliding window is used to detect the target contour boundary, and the phase data collected within the boundary is retained. The phase difference is used to calculate the propagation distance of the signal inside the target, and the tag coordinates and signal capture point coordinates are used to obtain the length of the signal direct path segment and AOA data. To obtain the theoretical phase of the propagation distance starting point, assume that the distance between the mth propagation distance starting point in direction 1 and its corresponding signal capture point is x m , using geometric relationships, we can obtain the distance between it and the label as The length of the i-th direct path segment is recorded as i∈[1,V x ] and i≠m, from the attached Figure 4 It can be found that there is a triangular relationship between the tag, the antenna moving position and the starting point of the propagation distance. The distance E between the propagation starting point and the signal capture point can be obtained by solving the triangular relationship based on the direct path segment length and AOA data. m , substitute into the ideal phase expression φ m =(4πE m / λ)mod 2π to obtain the phase value matrix of the propagation starting point Direction 2 can also obtain the matrix

[0028] Take label 1 as an example, as shown in the attached Figure 5 As shown, the position of label 1 (L / 2, 0) is taken as the pole O1 of the polar coordinate reference system, the positive direction of the horizontal coordinate of the Cartesian coordinate system is taken as the polar axis of the polar coordinate system, and the polar diameter r and polar angle are used. Represents the coordinates of each point in the imaging area in the polar coordinate system. The idea of ​​line segmentation is used to divide the direct signal path into several sub-segments, and the coordinates of the sub-segment are represented by the midpoint coordinates of the sub-segment. Using the formula Get the signal from The phase value measured when the light is reflected back through the sub-line segment T in Indicates the position of the midpoint Q of any sub-segment T and any signal capture point By analogy, we can obtain the theoretical phase value set of all direct path line segments and sub-line segments in two directions. Direction 2 can also obtain a collection Afterwards and Respectively and By comparison, we can obtain the probability that each sub-segment is the starting point of the propagation distance, and use the formula Calculate the pixel value for quantization, where T i，j Represents the jth sub-segment on the i-th direct path segment.

[0029] As attached Figure 5 As shown in the figure, the coordinates of the end point of the propagation distance can be obtained according to the coordinates of the starting point of the propagation distance, the propagation distance and the corresponding AOA data. The polar coordinates of the starting point and the end point of the propagation distance can be converted into Cartesian coordinates according to the coordinate conversion formula. The obtained starting point and the end point of the propagation distance are connected in sequence in different directions to form two closed figures connected by points and lines. The closed figures are filled with colors to obtain two estimated images. After merging, the ratio of the overlapping area to the total area is compared to obtain the overlap of the two estimated images. By adjusting the distance X between the starting point of the propagation distance and the signal capture point, the overlap of the estimated images in the two directions is maximized. The maximum number of iterations set is used as the termination condition, and the optimization model for the starting point of the propagation distance is constructed. The swarm optimization algorithm is introduced to solve it, and the fast and accurate search of the starting point and the end point of the propagation distance is realized.

[0030] Holographic positioning is performed on the optimal starting and ending points of the propagation distance in the plane grid area to obtain the pixel values ​​of the starting and ending points of the propagation distance in the plane grid area. The formula PX normalization =PX / PX max Normalize all pixel values ​​of a single target edge point, where PX represents the unnormalized pixel value matrix of a target edge point, PX maxRepresents the largest pixel value in the pixel value matrix, color maps the pixel values ​​of different sizes and fills the corresponding grid to form a hologram of a single target edge point, and uses the weight coefficient ω with a value range of 0 to 1 to suppress the pixel values ​​of the non-target edge point grid, establishes a set of target edge point holograms in two label directions, and superimposes the holograms contained in a single direction to obtain a single-direction sub-hologram of the target horizontal section. The sub-holograms in two directions are superimposed to generate a composite hologram of the target horizontal section. In order to better evaluate the imaging performance of the imaging system, a target imaging accuracy evaluation method based on deviation is proposed. The image deviation refers to the degree to which each point on the edge of the estimated image target deviates from the corresponding point on the actual target edge, that is, the positioning error of the estimated point is used as the imaging error, which is numerically reflected as the average value of the distance between the estimated point and the actual point. The deviation is an imaging accuracy evaluation index based on the target prior information. The smaller its value, the higher the positioning accuracy of the target edge point, and thus the higher the accuracy of the estimated image. For the estimated image of the imaging target, the imaging error is quantified, and the calculation formula of its deviation p is: Where (x′ n , y′ n ) represents the estimated coordinates of the nth point on the edge of the image; (x n ,y n ) represents the coordinates of the nth point on the actual target edge; N represents the number of located target edge points. Figure 7 The deviation curves of two arc-shaped targets are shown in the figure below. Figure 8 The deviation curves of two arc-shaped targets under three different noise environments are shown in the attached figure. Fig. 9 Comparison of the deviation curves of the imaging target under different line segment step sizes.

[0031] By the attached Figure 7 It can be seen that for target A with a circular horizontal section shape and a radius R = 15 cm, the minimum deviation is 0.1302 cm and the maximum deviation is 3.1897 cm; for target B with an elliptical horizontal section shape and a major axis 2a = 40 cm and a minor axis 2b = 20 cm, the minimum deviation is 0.2509 cm and the maximum deviation is 3.0218 cm, indicating that the target imaging system has good imaging performance for the two arc-shaped targets.

[0032] By the attached Figure 8 It can be seen that for target A, in several Gaussian white noise environments with different signal-to-noise ratios, when the number of iterations of the target imaging system reaches 50, the deviation curve begins to flatten, and the maximum deviation remains at about 2.5cm; for target B, in several Gaussian white noise environments with different signal-to-noise ratios, when the number of iterations reaches 50, the deviation curve also tends to flatten, and the maximum deviation remains at about 2.6cm. Fig. 9It can be seen that for the arc-shaped target with a radius of R = 18 cm, different step sizes are used for testing. As the number of iterations increases, the smaller the line segment step size, the smaller the deviation of the imaging target. This proves that reducing the line segment step size helps to improve the imaging accuracy of the system. Therefore, the effectiveness of the target imaging method proposed in the present invention can be proved. In particular, it avoids the adverse effects of imaging using the candidate target contour, and has a better imaging effect for arc-shaped targets. Compared with other imaging methods, it has better improvements in imaging efficiency and imaging accuracy.

Claims

1. An RFID target imaging method based on line segmentation and holographic positioning, characterized in that: The specific steps are as follows: Step 1: Build a system through RFID readers, antennas, tags and controllable mobile robots to collect tag phases and obtain the signal direct path propagation distance and angle of arrival (AOA) data. Establish a two-dimensional Cartesian coordinate system in the detection area with a length of L and a width of W, and set the controllable mobile robot to start from the initial position (0, W) at a speed v and move along a given trajectory. Set the coordinates of the two tags to (L / 2, 0) and (0, W / 2) respectively, define the vertical direction of tag T1 as direction 1, and define the horizontal direction of tag T2 as direction 2. Collect the tag phases in both the case of no target and target, and take the difference between the two to obtain two sets of phase differences Δφ. x and Δφ y The cumulative phase change method based on sliding window is used to detect the target contour boundary, and the phase data collected within the boundary is retained, which is recorded as Δφ′ x and Δφ′ y , the data volume is recorded as V x and V y , using the phase Δφ′ x and Δφ′ y Calculate the propagation distance PD of the signal inside the target x and PD y , the calculation formula is PD = (Δφ + 2σπ) / (β tar -β air ), where β air and β tar They represent the phase constants of the signal in the air and in the target respectively. Since the phase change caused by the signal penetrating the target generally does not exceed 2π, σ is usually 0. The signal capture point position AP can be obtained from the robot's moving speed and the timestamp when the signal is captured by the antenna. x and AP y , assuming and Respectively represent the coordinates of a signal capture point in direction 1 and direction 2. According to the formula Get the length DL of the direct path line segment of the signal in both directions x ,DL y , according to the formula Get AOA dataθ x ,θ y ; Step 2: Obtain the theoretical phase of the starting point of the propagation distance. Take the mth propagation distance starting point in direction 1 as an example. Assume that the distance between the mth propagation distance starting point and its corresponding signal capture point is x. m , then the distance between it and the label is The length of the i-th direct path segment is recorded as i∈[1,V x ] and i≠m, the angle between the i-th direct path segment and the direct path segment where the m-th propagation distance starting point is located is recorded as i∈[1,V x ] and i≠m, using the formula The distance between the starting point of the mth propagation distance and the other i signal capture points can be obtained as Then E m Substituting into the expression φ m =(4πE m / λ)mod 2π, the theoretical phase value φ at the starting point of the mth propagation distance under ideal conditions can be obtained. m , and so on, we can get the theoretical phase value matrix of all propagation distance starting points in two directions and Step 3: Construct a direct path segment decomposition model in the polar coordinate system, calculate the theoretical phase value of each direct path segment sub-segment, take the label 1 direction as an example, take the label 1 position (L / 2, 0) as the polar coordinate reference system pole O1, take the positive direction of the Cartesian coordinate system abscissa as the polar axis of the polar coordinate system, and use the polar diameter r and polar angle represents the coordinates of each point in the imaging area in the polar coordinate system, then each point on the direct path segment O1A can be To indicate that, It is obtained by converting the AOA data of the direct path segment to r f is the segment step length. The idea of ​​segment segmentation is used to segment O1A into several sub-segments. The coordinates of the midpoint of the sub-segment are the coordinates of the sub-segment. In the same polar coordinate system, the position A of the signal capture point x Can be used To represent the coordinates, where r j The length Dt of each direct path segment x Assume that the coordinates of the midpoint Q of any sub-segment T of O1A in the polar coordinate system are The location of any signal capture point The coordinates of The distance between the two You can use the formula Calculated, where And j∈[1,V x ], Then use the formula 2π can get the signal from The phase value measured when the light is reflected back through the sub-line segment T By analogy, the theoretical phase value set of all direct path line segments and sub-line segments in two directions can be obtained: and Step 4: Use the idea of ​​holographic positioning to complete the positioning of the starting point of the propagation distance. and Respectively and By comparison, we can obtain the probability that each sub-segment is the starting point of the propagation distance, and use the formula Calculate the pixel value for quantization, where T i，j Represents the jth sub-segment on the i-th direct path segment. For a direct path segment, find the sub-segment with the largest pixel value belonging to the direct path segment and the coordinates of the starting point of the propagation distance It is the coordinate of the midpoint of the sub-segment with the largest pixel value, and the same is true for other direct path segments; Step 5: Construct an optimization model based on maximizing overlap to find the optimal starting point of the propagation distance. The coordinates of the end point of the propagation distance can be obtained based on the coordinates of the starting point of the propagation distance, the propagation distance, and the corresponding AOA data. The calculation formula for the coordinates of the end point of the propagation distance in the tag 1 direction is: Among them, i∈[1,V x ], The calculation formula for the coordinates of the end point of the propagation distance in the tag 2 direction is: Among them, i∈[1,V y ], Convert the polar coordinates of the starting and ending points of the propagation distance into Cartesian coordinates. The conversion formula for label 1 direction is: The formula for label 2 direction conversion is: The starting point and the end point of the propagation distance are connected in sequence in different directions to form two closed figures connected by points and lines. The closed figures are filled with colors to obtain two estimated images. After merging, the ratio of the overlapping area to the total area is compared to obtain the overlap of the two estimated images. The overlap of the estimated images in the two directions is maximized by adjusting the distance X between the starting point of the propagation distance and the signal capture point. The maximum number of iterations set is used as the termination condition to construct an optimization model for the starting point of the propagation distance. The swarm optimization algorithm is introduced to solve the problem, so as to realize the fast and accurate search of the starting point and the end point of the propagation distance. Step 6: Aggregate all the optimal propagation distance starting points and ending points into a hologram to generate a composite hologram. Use the composite hologram reconstruction method based on the target prior information to present the horizontal section image of the target. According to the imaging scene deployment and data acquisition method, perform holographic positioning of the optimal propagation distance starting point and ending point in the plane grid area to obtain the pixel values ​​of the propagation distance starting point and ending point in the plane grid area. All the propagation distance starting points and ending points are collectively referred to as target edge points. Use the formula PX normalization =PX / PX max Normalize all pixel values ​​of a single target edge point, where PX represents the unnormalized pixel value matrix of a target edge point, PX max Represents the largest pixel value in the pixel value matrix, performs color mapping on pixel values ​​of different sizes and fills the corresponding grids to form a hologram of a single target edge point, uses a weight coefficient ω ranging from 0 to 1 to suppress the pixel values ​​of the non-target edge point grid, establishes a set of target edge point holograms in two label directions, superimposes the holograms contained in a single direction, obtains a single-direction sub-hologram of the target horizontal section, and superimposes the sub-holograms in two directions to generate a composite hologram of the target horizontal section.

2. The method according to claim 1, characterized in that In step 6, the basis for holographic positioning of the starting point and the end point of the propagation distance is that the sub-segments on each direct path segment belong to different angular directions, and when the same segment step size is used to divide each direct path segment, the number of sub-segments in each direction is not exactly the same, so the target edge point hologram based on segment division cannot be superimposed. In addition, all target edge point holograms based on segment division have no intersection in space, and simply superimposing them cannot reconstruct the horizontal section image of the target. Therefore, all target edge points need to be holographically positioned to obtain their pixel values ​​in the plane grid area, and then generate a hologram based on plane division to complete image reconstruction.

Images