Lightweight discrimination system and method for RFID tag and point cloud cargo association

Abstract

The invention discloses a lightweight discrimination system and method for association of RFID tags and point cloud goods, and belongs to the technical field of fusion of RFID wireless sensing and environmental sensing. The implementation method comprises the following steps: 1, constructing a mobile platform, and setting a track for the mobile platform; 2, obtaining an observation data set, and taking a position coordinate of the mobile platform at a signal intensity peak moment as a physical distance between the ith RFID tag and the mobile platform; 3, acquiring a signal intensity value, and acquiring the distance from the real spatial position Li of the to-be-observed label to the spatial position of the mobile platform; 4, geometric features of the point cloud object blocks are obtained; 5, constraining the point cloud object block through spatial constraint, direction constraint and scoring constraint, and obtaining an RFID tag corresponding to the point cloud object block; and 6, constructing a mapping relation between the to-be-observed label and the point cloud object block, and outputting the mapping relation in a visual mode. Compared with the prior art, the technical problem of preventing missing detection of point cloud goods corresponding to the RFID tag in intelligent warehousing is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to a lightweight discrimination system and method for associating RFID tags with point cloud goods, belonging to the technical field of RFID wireless sensing and environmental sensing fusion, and applied to the aspects of material clearing and collaborative distribution of intelligent warehousing. BACKGROUND

[0002] In the prior art, the positioning and environment mapping scheme of mobile platforms in the fields of intelligent logistics, warehouse automation and emergency rescue mainly relies on SLAM technology. Through the multi-sensor fusion method of laser radar, visual camera and inertial measurement unit (IMU), a three-dimensional geometric model of the environment is constructed, and the pose of the robot is estimated in real time, realizing real-time perception and autonomous navigation of the environment. This method performs excellently in structured environments and has been widely applied in industrial production, warehouse management and other fields.

[0003] The existing intelligent warehousing currently has the following defects: 1. Unstable RFID signal: The signal strength (RSSI) of the RFID tag is easily affected by multipath effect, obstacles and environmental interference, resulting in unstable signal transmission, and further affecting the accurate positioning of the tag. 2. Inaccurate relationship between tag and object: The existing technology usually relies on RSSI signal strength to infer the spatial relationship between the tag and the object, but this method is prone to errors in dynamic environments, resulting in missed detection or false detection. 3. Low fusion efficiency of point cloud and RFID: Although point cloud technology can provide three-dimensional spatial information of objects, effective fusion of point cloud data and RFID tags still faces great challenges. The existing method usually requires complex algorithms and additional calibration steps, which is large in calculation amount and poor in real-time performance, and is difficult to cope with complex and dynamic warehouse environments.

[0004] Therefore, how to prevent the missed detection of point cloud goods corresponding to RFID tags in intelligent warehousing has become a problem to be solved. SUMMARY

[0005] The purpose of the present application is to solve the technical problem of preventing the missed detection of point cloud goods corresponding to RFID tags in intelligent warehousing, and a lightweight discrimination system and method for associating RFID tags with point cloud goods are proposed.

[0006] The working principle of the application is: the semantic attribution relationship model between the RFID tag and the three-dimensional object block is established by using the Bayesian peak inference of the received signal strength indication (RSSI) time sequence of the RFID tag and the point cloud display surface geometry matching. The trajectory point of the reader corresponding to the RSSI peak is taken as the rough observation basis of the tag, the object block contour and the visible surface orientation information obtained by the instance segmentation of the point cloud are combined, the display surface area constraint and the normalized distance scoring mechanism are adopted, the object block to which the tag is most likely attached is determined, and the tag information semantics is hung in the object block, so that the spatial attribution relationship between the tag and the object is visually expressed.

[0007] The purpose of the application is realized by the following technical solutions:

[0008] The application discloses a lightweight discrimination method for RFID tag and point cloud cargo association, which comprises the following steps:

[0009] Step 1: a mobile platform for identifying the shape information of the RFID tag and the point cloud cargo is constructed, and a trajectory for segmentally sensing the storage position of the point cloud cargo is set for the mobile platform;

[0010] Step 1.1: a mobile platform composed of an RFID transceiver, an antenna and a laser radar is constructed; wherein the RFID transceiver and the antenna are used for identifying the RFID tag; and the laser radar is used for identifying the shape information of the point cloud cargo;

[0011] Step 1.2: a motion trajectory for segmentally sensing the storage position of the point cloud cargo is set;

[0012] Step 2: the signal strength value of the i-th RFID tag of the cargo in the region of the k-th path and the point cloud frame data are obtained to form an observation data set, and the position coordinates of the mobile platform at the signal strength peak time are taken as the physical distance between the i-th RFID tag and the mobile platform;

[0013] Step 2.1: all point cloud cargo storage positions are identified by using the laser radar, and the motion trajectory is segmented by using the point cloud density and the storage position in the manner shown in formula (1);

[0014]

[0015] Wherein, K represents the total number of trajectory segments; represents the k-th path;

[0016] Step 2.2: the signal strength value of the i-th RFID tag of the cargo in the region of the k-th path at the current time t and the point cloud frame data are obtained by using the mobile platform to form an observation data set;

[0017] Step 2.2.1: Obtain the signal strength value of the i-th RFID tag of the goods in the area of the k-th path at the current time t by using the mobile platform;

[0018] Step 2.2.1.1: Set the frequency threshold for the mobile platform to collect the RFID tags; collect the RFID tags by using the frequency threshold;

[0019] Step 2.2.1.2: Obtain the signal strength value of the RFID tag by the distance between the RFID transceiver and the goods RFID tag;

[0020] Step 2.2.1.3: Obtain the signal strength value of the i-th RFID tag at the current time t as shown in formula (2);

[0021]

[0022] Wherein, RSSI(·) represents the signal strength value of the RFID tag;

[0023] Step 2.2.2: Obtain the point cloud frame data of the goods in the area of the k-th path at the current time t as shown in formula (3) by using the laser radar;

[0024]

[0025] Wherein, p(·) represents the point cloud frame data;

[0026] Step 2.2.3: Construct the observation data set as shown in formula (4) by using the signal strength value of the i-th RFID tag of the goods in the area of the k-th path and the point cloud frame data;

[0027]

[0028] Step 2.3: Take the position coordinates of the mobile platform at the signal strength peak value time of the i-th RFID tag of the goods in the area of the k-th path as the physical distance between the i-th RFID tag and the mobile platform;

[0029] Step 2.3.1: Obtain the signal strength peak value of the i-th RFID tag;

[0030] Step 2.3.2: Obtain the peak time corresponding to the signal strength peak value as shown in formula (5);

[0031]

[0032] Step 2.3.3: Obtain the position coordinates of the mobile platform corresponding to the peak time as shown in formula (6), which is taken as the physical distance between the i-th RFID tag and the mobile platform;

[0033]

[0034] wherein (x, y, z) represents the position coordinates of the mobile platform;

[0035] Step 3: Obtain the time t of the goods in the region of the kth path i Observe the signal strength value of the tag, and obtain the tag to be observed Real space position L i Distance to the space position of the mobile platform;

[0036] Step 3.1: Obtain the time t of the goods in the region of the kth path as shown in formula (7) i The set of perceived RFID tags;

[0037]

[0038] Step 3.2: Set the tag to be observed Real space position L i = [x Li , y Li , z Li ] T , the space position of the mobile platform at time t i

[0039] Step 3.3: Obtain the Euclidean distance between the mobile platform and the tag to be observed at time t i as shown in formula (8);

[0040]

[0041] Step 3.4: Obtain the tag to be observed by using the signal strength value of the tag to be observed at time t i as shown in formula (9) Real space position L i Distance to the space position of the mobile platform;

[0042]

[0043] Wherein P0 is the received signal strength at the reference distance (unit dBm), and n is the path loss index; is the observation noise or slight drift, satisfying

[0044] Step 4: Obtain the geometric features of the point cloud object block at time t i of the goods in the region of the kth path;

[0045] Step 4.1: Obtain the perceived point cloud time series at time t i of the goods in the region of the kth path as shown in formula (10)​

[0046]

[0047] Step 4.2: Use clustering segmentation method to extract the candidate point cloud object block set of the point cloud time series as shown in formula (11);

[0048]

[0049] Step 4.3: Get the centroid position Display surface outline, normal vector between display surface and mobile platform and the enclosing radius The geometric characteristics of the point cloud object blocks;

[0050] Step 5: Using the geometric features of the point cloud object block, constrain the point cloud object block through spatial constraints, direction constraints, and scoring constraints to obtain the RFID tag corresponding to the point cloud object block;

[0051] Step 5.1: Using the observed labels RSSI reference point position corresponding to each frame of point cloud

[0052] Step 5.2: With a peak moment of one frame Point cloud object blocks in Make a comparison;

[0053] Step 5.3: Constrain the point cloud object block using spatial constraints, directional constraints, and scoring constraints to obtain the RFID tag corresponding to the point cloud object block;

[0054] Step 5.3.1: Acquisition using spatial constraints The display surface outline;

[0055] Step 5.3.2: Use the orientation constraint to display the normal vectors of the surface and the mobile platform

[0056] Step 5.3.3: Use the scoring constraint shown in Equation (12) to fuse the display surface contour and normal vector to obtain the RFID tag corresponding to the point cloud object block;

[0057]

[0058] Where ε is the smoothing constant;

[0059] Step 6: Construct a mapping relationship between the observed label and the point cloud object block, and output it in a visual way;

[0060] Step 6.1: Construct the mapping relationship between the observed label and the point cloud object block as shown in formula (13);

[0061]

[0062] Wherein, RFID_id i is a unique identifier; is a point cloud object block number; relative_xy i is a two-dimensional local coordinate of the label to be observed on the display surface, used for visualization and semantic processing; remarks is a remark information;

[0063] Step 6.2: output the mapping relationship between the point cloud object block and the RFID label in a visual manner;

[0064] On the other hand, in order to achieve the purpose of the present application, according to the above method, the present application further proposes a lightweight discrimination system for associating RFID labels with point cloud goods, comprising a signal acquisition and processing module, a point cloud object block segmentation module and a label-object association reasoning module.

[0065] The signal acquisition and processing module is used to collect the RSSI signal of the RFID label at a fixed frequency during the movement of the mobile platform, construct a signal strength time sequence, and extract the peak signal observation point of each label as the rough spatial position observation point of the label through Bayesian peak reasoning, output the rough observation position data of the label as the input of the label-object association reasoning module.

[0066] The point cloud object block segmentation module is used to obtain the environmental point cloud data at each sampling time of the mobile platform, perform clustering or instance segmentation, extract the object block set, and extract the centroid position, main display surface geometric contour and normal vector of each object block, output the point cloud object block feature data as the input of the label-object association reasoning module.

[0067] The label-object association reasoning module is used to perform spatial constraint matching and display surface direction constraint matching of the label and the object based on the RSSI peak position data and the object block feature data, and determine the optimal attribution relationship between the label and the object block by combining the normalized distance scoring mechanism, and output the semantic mapping result of the label and the object block.

[0068] Advantages:

[0069] Compared with the prior art, the present application has the following advantages:

[0070] ​1. In terms of the accuracy of the association between the tag and the object, the present application realizes the efficient and accurate association between the tag and the object through the fusion matching method of the RSSI peak position and the point cloud object block display surface feature. Compared with the method relying only on single factors such as signal strength or geometric position, the present application significantly improves the accuracy and robustness of the tag space attribution judgment.

[0071] 2. In terms of semantic information visualization, the present application directly mounts the tag information to the display surface area of the point cloud object block, forming a clear and intuitive semantic spatial relationship, and improving the interactivity and maintainability of the material management system for the on-site environment.

[0072] 3. The method of the present application can complete the spatial unified expression of the tag and the point cloud data without complex system calibration or additional sensor fusion processing, greatly reducing the deployment cost and operation complexity, and being suitable for various logistics and warehousing scenes.

[0073] 4. In terms of adaptability in complex scenes, the present application effectively controls signal noise and observation error using a Bayesian inference model, effectively filters misjudgment situations through a multi-constraint mechanism, and enhances the operation reliability and stability of the system in the presence of signal interference, shielding and unstructured environment. BRIEF DESCRIPTION OF DRAWINGS

[0074] Figure 1 is a flowchart of the present application;

[0075] Figure 2 is an experimental result visualization interface diagram of the present application. DETAILED DESCRIPTION

[0076] In order to better illustrate the purpose and advantages of the present application, the following further describes the content of the application in combination with the drawings and examples. It should be noted that the implementation of the present application is not limited to the following examples, and any form of variation or change of the present application will fall within the scope of protection of the present application.

[0077] EMBODIMENT

[0078] As shown in Figure 1 , a lightweight discrimination method for associating RFID tags with point cloud goods of the present embodiment is as follows:

[0079] Step 1: Construct a mobile platform for identifying the shape information of RFID tags and point cloud goods, and set a trajectory for the mobile platform to segmentally perceive the storage position of the point cloud goods;

[0080] Step 1.1: Construct a mobile platform composed of an RFID transceiver, an antenna and a laser radar; wherein the RFID transceiver and the antenna are used to identify the RFID tag; the laser radar is used to identify the shape information of the point cloud goods;

[0081] In the embodiment, the working frequency of the RFID tag is in the range of 902-928 MHz, and the laser radar adopts a multi-line solid-state laser radar.

[0082] Step 1.2: Set the motion trajectory for segmenting the point cloud of the sensed goods storage location;

[0083] Step 2: Obtain the signal strength value of the i-th RFID tag in the region of the k-th path and the point cloud frame data to form an observation data set, and take the position coordinate of the mobile platform at the signal strength peak time as the physical distance between the i-th RFID tag and the mobile platform;

[0084] Step 2.1: Identify all point clouds of the goods storage location using the laser radar, and segment the motion trajectory using the point cloud density and the storage location in the manner shown in equation (1);

[0085]

[0086] wherein K represents the total number of trajectory segments; represents the k-th path;

[0087] In the embodiment, the total number of trajectory segments K = 2; the path represents the goods storage area 1, and the path represents the goods storage area 2.

[0088] Step 2.2: Obtain the signal strength value of the i-th RFID tag of the goods in the region of the k-th path at the current time t using the mobile platform, and form an observation data set;

[0089] Step 2.2.1: Obtain the signal strength value of the i-th RFID tag of the goods in the region of the k-th path at the current time t using the mobile platform;

[0090] Step 2.2.1.1: Set the frequency threshold for the mobile platform to collect the RFID tag; and collect the RFID tag using the frequency threshold;

[0091] Step 2.2.1.2: Obtain the signal strength value of the RFID tag by the distance between the RFID transceiver and the goods RFID tag;

[0092] Step 2.2.1.3: Obtain the signal strength value of the i-th RFID tag at the current time t as shown in equation (2);

[0093]

[0094] wherein RSSI(·) represents the signal strength value of the RFID tag;

[0095] In the embodiment, the path is in the cargo storage area 1, i.e. k = 1; the signal strength value is -40 dBm;

[0096] Step 2.2.2: The laser radar used acquires the point cloud frame data of the cargo in the region of the kth path at the current time t as shown in formula (3);

[0097]

[0098] Wherein, p(·) represents the point cloud frame data;

[0099] Step 2.2.3: The signal strength value of the ith RFID tag of the region cargo of the kth path and the point cloud frame data are constructed into an observation data set as shown in formula (4);

[0100]

[0101] Step 2.3: The position coordinates of the mobile platform at the signal strength peak value time of the ith RFID tag in the region of the kth path are taken as the physical distance between the ith RFID tag and the mobile platform;

[0102] Step 2.3.1: The signal strength peak value of the ith RFID tag is acquired;

[0103] Step 2.3.2: The peak time corresponding to the signal strength peak value as shown in formula (5) is acquired;

[0104]

[0105] Step 2.3.3: The position coordinates of the mobile platform corresponding to the peak time as shown in formula (6) are acquired, which are taken as the physical distance between the ith RFID tag and the mobile platform;

[0106]

[0107] Wherein, (x, y, z) represents the position coordinates of the mobile platform;

[0108] Step 3: The time t i The observed tag signal strength value is acquired, and the observed tag The real space position L i The distance to the mobile platform space position;

[0109] Step 3.1: The time t i of the perception of the set of RFID tags in the region of the kth path as shown in formula (7) is acquired;

[0110]

[0111] Step 3.2: Set the label to be observed Real space position L i =[x Li ,y Li ,z Li ] T , mobile platform in t i Time and space position

[0112] Step 3.3: Obtain the time t as shown in formula (8) i The Euclidean distance between the mobile platform and the observed label;

[0113]

[0114] Step 3.4: Use the method as in formula (9) and use the time t i The signal strength value of the tag to be observed is used to obtain the tag to be observed Real space position L i The distance to the spatial location of the mobile platform;

[0115]

[0116] Where P0 is the received signal strength at the reference distance (in dBm), and n is the path loss exponent; To observe noise or small drift, satisfy

[0117] In the embodiment, is -40dBm, P0 is -30dBm, n is 2, is 1, About 3.55m;

[0118] Step 4: Get the time t of the goods in the area of ​​the kth path i The geometric features of the point cloud object blocks;

[0119] Step 4.1: Obtain the time t of the goods in the area of ​​the kth path as shown in formula (10) i The perceived point cloud time series;

[0120]

[0121] Step 4.2: Use clustering segmentation method to extract the candidate point cloud object block set of the point cloud time series as shown in formula (11);

[0122]

[0123] Step 4.3: Get the centroid position Display surface outline, normal vector between display surface and mobile platform and the enclosing radius r i (t) geometric features of the point cloud object block;

[0124] In an embodiment, the enclosing radius r i (t) is 0.5m;

[0125] Step 5: using the geometric features of the point cloud object block, the point cloud object block is constrained by spatial constraint, direction constraint and scoring constraint, and the RFID tag corresponding to the point cloud object block is obtained;

[0126] Step 5.1: using the to-be-observed tag RSSI reference point position corresponding to each frame of point cloud

[0127] Step 5.2: comparing with the point cloud object block in the peak moment of a frame;

[0128] Step 5.3: using the spatial constraint, the direction constraint and the scoring constraint to constrain the point cloud object block, and obtaining the RFID tag corresponding to the point cloud object block;

[0129] Step 5.3.1: using the spatial constraint to obtain the display surface contour of

[0130] Step 5.3.2: using the direction constraint to fuse the normal vector of the display surface and the moving platform

[0131] Step 5.3.3: using the scoring constraint as shown in formula (12) to fuse the display surface contour and the normal vector, and obtaining the RFID tag corresponding to the point cloud object block;

[0132]

[0133] wherein, ε is a smoothing constant;

[0134] In an embodiment, ε is 0.01;

[0135] Step 6: constructing the mapping relationship between the to-be-observed tag and the point cloud object block, and outputting by visualizing;

[0136] Step 6.1: constructing the mapping relationship between the to-be-observed tag and the point cloud object block as shown in formula (13);

[0137]

[0138] wherein, RFID_id i is​​ the unique identifier of the RFID tag; the block number of the point cloud object; relative_xy i the two-dimensional local coordinates of the to-be-observed label on the display surface, used for visualization and semantic processing; and remarks, remark information;

[0139] Step 6.2: output the mapping relationship between the point cloud object block and the RFID tag in a visual manner;

[0140] In the embodiment, as shown in FIG. 6, the experimental result is visualized and output. Figure 2

[0141] On the other hand, in order to achieve the object of the present application, according to the above method, the present application further provides a lightweight discrimination system for associating RFID tags with point cloud goods, comprising a signal acquisition and processing module, a point cloud object block segmentation module, and a label-object association reasoning module.

[0142] The signal acquisition and processing module is configured to acquire the RSSI signal of the RFID tag at a fixed frequency during the movement of the mobile platform, construct a signal strength time sequence, and extract the peak signal observation point of each label as the rough spatial position observation point of the label through Bayesian peak reasoning, and output the rough observation position data of the label as the input of the label-object association reasoning module.

[0143] The point cloud object block segmentation module is configured to acquire the environmental point cloud data at each sampling time of the mobile platform, perform clustering or instance segmentation, extract a set of object blocks, and extract the centroid position, main display surface geometric contour, and normal vector of each object block, and output the point cloud object block feature data as the input of the label-object association reasoning module.

[0144] The label-object association reasoning module is configured to perform spatial constraint matching and display surface direction constraint matching of the label and the object based on the RSSI peak position data and the object block feature data, and determine the optimal attribution relationship between the label and the object block in combination with a normalized distance scoring mechanism, and output the semantic mapping result of the label and the object block.

[0145] The above detailed description further describes the object, technical solution, and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.​

Claims

1. A lightweight method for determining the association between RFID tags and point cloud goods, characterized by: The following steps are included: Step 1: Build a mobile platform for identifying the shape information of RFID tags and point cloud goods, and set a trajectory for the mobile platform to segmentally perceive the storage location of point cloud goods; Step 2: Obtain the signal strength value and point cloud frame data of the i-th RFID tag of the goods in the area of ​​the k-th path to form an observation dataset, and use the position coordinates of the mobile platform at the moment of signal strength peak as the physical distance between the i-th RFID tag and the mobile platform; Step 3: Get the time t of the goods in the area of ​​the kth path i The signal strength value of the tag to be observed and the tag to be observed are obtained Real space position L i The distance to the spatial location of the mobile platform; Step 4: Get the time t of the goods in the area of ​​the kth path i The geometric features of the point cloud object blocks; Step 5: Using the geometric features of the point cloud object block, constrain the point cloud object block through spatial constraints, direction constraints, and scoring constraints to obtain the RFID tag corresponding to the point cloud object block; Step 5.1: Using the observed labels RSSI reference point position corresponding to each frame of point cloud Step 5.2: With a frame peak moment Point cloud object blocks in Make a comparison; Step 5.3: Constrain the point cloud object block using spatial constraints, directional constraints, and scoring constraints to obtain the RFID tag corresponding to the point cloud object block; Step 5.3.1: Acquisition using spatial constraints The display surface outline; Step 5.3.2: Use the orientation constraint to display the normal vectors of the surface and the mobile platform Step 5.3.3: Use the scoring constraint shown in Equation (12) to fuse the display surface contour and normal vector to obtain the RFID tag corresponding to the point cloud object block; Where ε is the smoothing constant; Step 6: Construct the mapping relationship between the observed label and the point cloud object block, and output it in a visual way.

2. The lightweight method for determining the association between an RFID tag and a point cloud item according to claim 1, wherein: Step 1 is implemented as follows: Step 1.1: Build a mobile platform consisting of an RFID transceiver, antenna, and lidar. The RFID transceiver and antenna are used to identify RFID tags, while the lidar is used to identify the shape information of the point cloud goods. Step 1.2: Set the motion trajectory for segmented perception of the point cloud cargo storage location.

3. The lightweight method for determining the association between an RFID tag and a point cloud item according to claim 1, wherein: Step 2 is implemented as follows: Step 2.1: Use LiDAR to identify the storage locations of all point clouds and segment the trajectory using the method shown in formula (1) based on the point cloud density and storage location; Where K represents the total number of trajectory segments; represents the kth path; Step 2.2: Use the mobile platform to obtain the signal strength value and point cloud frame data of the i-th RFID tag of the goods in the area of ​​the k-th path at the current time t to form an observation dataset; Step 2.3: The position coordinates of the mobile platform at the peak moment of the signal strength of the i-th RFID tag of the goods in the area of ​​the k-th path are used as the physical distance between the i-th RFID tag and the mobile platform.

4. The lightweight method for determining the association between an RFID tag and a point cloud item according to claim 3, wherein: The implementation method of step 2.2 is: Step 2.2.1: Use the mobile platform to obtain the signal strength value of the i-th RFID tag of the goods in the area of ​​the k-th path at the current time t; Step 2.2.2: Use the laser radar to obtain the point cloud frame data of the cargo in the area of ​​the k-th path at the current time t as shown in formula (3); Where p(·) represents the point cloud frame data; Step 2.2.3: Construct an observation dataset as shown in formula (4) using the signal strength value and point cloud frame data of the i-th RFID tag of the regional goods in the k-th path.

5. The lightweight method for determining the association between an RFID tag and a point cloud item according to claim 3, wherein: The implementation method of step 2.3 is: Step 2.3.1: Obtain the peak signal strength of the i-th RFID tag; Step 2.3.2: Obtain the peak time corresponding to the signal strength peak as shown in formula (5); Step 2.3.3: Obtain the position coordinates of the mobile platform corresponding to the peak moment as shown in formula (6), which will be used as the physical distance between the i-th RFID tag and the mobile platform; Among them, (x, y, z) represents the position coordinates of the mobile platform.

6. The lightweight method for determining the association between an RFID tag and a point cloud item according to claim 4, wherein: The implementation method of step 2.2.1 is: Step 2.2.1.1: Set the frequency threshold for the mobile platform to collect RFID tags; use the frequency threshold to collect RFID tags; Step 2.2.1.2: Obtain the signal strength value of the RFID tag based on the distance between the RFID transceiver and the RFID tag of the goods; Step 2.2.1.3: Obtain the signal strength value of the i-th RFID tag at the current time t as shown in formula (2); Wherein, RSSI(·) represents the signal strength value of the RFID tag.

7. The lightweight method for determining the association between an RFID tag and a point cloud item according to claim 1, wherein: Step 3 is implemented as follows: Step 3.1: Obtain the time t of the goods in the area of ​​the kth path as shown in formula (7) i A collection of sensing RFID tags; Step 3.2: Set the label to be observed Real space position L i =[x Li ,y Li ,z Li ] T , mobile platform in t i Time and space position Step 3.3: Obtain the time t as shown in formula (8) i The Euclidean distance between the mobile platform and the observed label; Step 3.4: Use the method as in formula (9) and use the time t i The signal strength value of the tag to be observed is used to obtain the tag to be observed Real space position L i The distance to the spatial location of the mobile platform; Where P0 is the received signal strength at the reference distance (in dBm), and n is the path loss exponent; To observe noise or small drift, satisfy 8. The lightweight method for determining the association between an RFID tag and a point cloud item according to claim 1, wherein: Step 4 is implemented as follows: Step 4.1: Obtain the time t of the goods in the area of ​​the kth path as shown in formula (10) i The perceived point cloud time series; Step 4.2: Use clustering segmentation method to extract the candidate point cloud object block set of the point cloud time series as shown in formula (11); Step 4.3: Get the centroid position Display surface outline, normal vector between display surface and mobile platform and the enclosing radius r i (t) The geometric features of the point cloud object blocks.

9. The lightweight method for determining the association between an RFID tag and a point cloud item according to claim 1, wherein: Step 6 is implemented as follows: Step 6.1: Construct the mapping relationship between the observed label and the point cloud object block as shown in formula (13); Among them, RFID_id i for unique identifier; Number the point cloud object blocks; relative_xy i The two-dimensional local coordinates of the observed label on the display surface are used for visualization and semantic processing; remarks are remarks; Step 6.2: Output the point cloud object blocks and RFID tags in the mapping relationship in a visual manner.

10. A lightweight method for determining the association between an RFID tag and a point cloud product according to claim 1, characterized in that: It includes signal acquisition and processing module, point cloud object block segmentation module and label object association reasoning module; The signal acquisition and processing module is used to collect RSSI signals of RFID tags at a fixed frequency during the movement of the mobile platform, construct a signal strength time series, and extract the peak signal observation point of each tag as the rough spatial position observation point of the tag through Bayesian peak inference, and output the rough observation position data of the tag as the input of the tag-object association inference module; The point cloud object block segmentation module is used to acquire environmental point cloud data at each sampling moment of the mobile platform, perform clustering or instance segmentation, extract a set of object blocks, and extract features such as the center of mass position, main display surface geometric outline, and normal vector of each object block, and output point cloud object block feature data as input to the label object association inference module; The label-object association inference module is used to perform spatial constraint matching and display surface direction constraint matching between labels and objects based on RSSI peak position data and object block feature data, and combine the normalized distance scoring mechanism to determine the optimal attribution relationship between labels and object blocks, and output the semantic mapping results between labels and object blocks.

Images