Warehouse goods high-precision positioning and automatic inventorying method, system and equipment

Abstract

The application discloses a warehouse goods high-precision positioning and automatic inventory method, system and equipment, and relates to the field of warehouse management.The method comprises the following steps: constructing a fingerprint library; the fingerprint library comprises the received signal strength of each reference tag; the reference tag is a passive radio frequency identification tag arranged on a shelf in a warehouse; a scanning point is generated and a flight path is planned according to an inventory task and the layout of the shelves in the warehouse; a machine vision algorithm and a radio frequency identification positioning method are used to control a UAV to pass through the scanning points according to the flight path; the received signal strength of a to-be-measured tag is collected by a radio frequency identification reader carried by the UAV; the to-be-measured tag is a passive radio frequency identification tag arranged on goods in the warehouse; the spatial position of the to-be-measured tag is calculated by using an iterative Bayes-based radio frequency identification fingerprint algorithm according to the received signal strength of the to-be-measured tag and the fingerprint library, so that the goods in the warehouse are inventoried. The application improves the inventory accuracy and efficiency of the warehouse and reduces the hardware deployment cost.

Description

Technical Field

[0001] This application relates to the field of warehouse management, and in particular to a method, system and equipment for high-precision positioning and automatic inventory counting of warehouse goods. Background Technology

[0002] Inventory counting is a crucial part of modern warehouse management. Regularly checking finished goods and raw materials in the warehouse ensures that the goods are in good condition and that the quantities and types match the records, thus guaranteeing the smooth operation of production and the flow of goods. However, with socio-economic development, the volume and throughput of warehouse goods are constantly increasing. Traditional methods based on paper lists and manual inventory counting are time-consuming, inefficient, and prone to omissions, becoming obstacles to production and the flow of goods.

[0003] Drones equipped with sensors can move and explore quickly in warehouses. To optimize warehouse management efficiency, many intelligent inventory management solutions based on drones have emerged. The most important aspect is the choice of positioning technology. Common indoor positioning technologies such as Wireless Fidelity (WIFI), Bluetooth, and Long Term Evolution (LTE) / 5G have low positioning accuracy. Acoustic and ultrasonic technologies require strict control of ambient temperature and humidity. Ultra-wideband (UWB) and Channel State Information (CSI) have special requirements for the performance of transceivers and signal modulation. These solutions either have positioning accuracy that is difficult to meet the needs of warehouse management or require high equipment and environmental modification costs, which restricts the implementation of intelligent solutions. Summary of the Invention

[0004] The purpose of this application is to provide a method, system, and equipment for high-precision positioning and automatic inventory counting of warehouse goods, which can improve positioning accuracy and inventory counting efficiency, and reduce costs.

[0005] To achieve the above objectives, this application provides the following solution:

[0006] Firstly, this application provides a method for high-precision positioning and automatic inventory counting of warehouse goods, including:

[0007] Construct a fingerprint database; the fingerprint database includes the received signal strength of each reference tag; the reference tag is a passive radio frequency identification tag deployed on the shelves of the warehouse;

[0008] Based on the inventory task and the layout of the shelves in the warehouse, scan points are generated and flight paths are planned;

[0009] The unmanned aerial vehicle is controlled to pass through the scanning points according to the flight path by using a machine vision algorithm and a radio frequency identification positioning method, and meanwhile, a radio frequency identification reader carried by the unmanned aerial vehicle collects the received signal strength of the to-be-tested label; the to-be-tested label is a passive radio frequency identification label arranged on goods in the warehouse;

[0010] According to the received signal strength of the to-be-tested label and the fingerprint library, a radio frequency identification fingerprint algorithm based on iterative Bayes is used to solve the spatial position of the to-be-tested label, so as to count the goods in the warehouse.

[0011] In a second aspect, the present application provides a warehouse goods high-precision positioning and automatic counting system, comprising:

[0012] A fingerprint library construction module is configured to construct a fingerprint library; the fingerprint library includes the received signal strength of each reference label; the reference label is a passive radio frequency identification label arranged on a goods shelf in the warehouse;

[0013] A path planning module is configured to generate scanning points and plan a flight path according to a counting task and the layout of the goods shelves in the warehouse;

[0014] An unmanned aerial vehicle control module is configured to control the unmanned aerial vehicle to pass through the scanning points according to the flight path by using a machine vision algorithm and a radio frequency identification positioning method; in the process of flight of the unmanned aerial vehicle, a radio frequency identification reader carried by the unmanned aerial vehicle collects the received signal strength of the to-be-tested label; the to-be-tested label is a passive radio frequency identification label arranged on goods in the warehouse;

[0015] A position solving module is configured to acquire the received signal strength of the to-be-tested label collected by the unmanned aerial vehicle, and according to the received signal strength of the to-be-tested label and the fingerprint library, a radio frequency identification fingerprint algorithm based on iterative Bayes is used to solve the spatial position of the to-be-tested label, so as to count the goods in the warehouse.

[0016] In a third aspect, the present application provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the warehouse goods high-precision positioning and automatic counting method described above.

[0017] According to the specific embodiments provided by the present application, the present application has the following technical effects:

[0018] The warehouse goods high-precision positioning and automatic inventory method, system and equipment provided by the application adopt VINS-Mono machine vision technology and a Radio Frequency Identification (RFID) positioning method to position a UAV, improve the positioning accuracy of the UAV, adopt an RFID fingerprint algorithm based on iterative Bayes to position goods, improve the positioning accuracy of the goods, and further improve the accuracy of warehouse inventory. When inventorying, scanning points are generated and flight paths are planned according to inventory tasks and the layout of shelves in the warehouse, improving the inventory efficiency. The entire scheme only needs to set passive RFID tags in the warehouse and carry RFID readers on the UAV, reducing the hardware deployment cost. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0020] Figure 1 The overall flowchart of a warehouse goods high-precision positioning and automatic inventory method provided by an embodiment of the present application.

[0021] Figure 2 The detailed flowchart of a warehouse goods high-precision positioning and automatic inventory method provided by an embodiment of the present application.

[0022] Figure 3 The overall structure diagram of a warehouse goods high-precision positioning and automatic inventory system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0024] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0025] From the perspective of intelligence, there are currently the following three means for warehouse goods inventory.

[0026] (1) Manual inventory, manual visual acquisition of material information and verification with paper material list one by one, then manual input into the system, the hardware cost is the lowest, but the efficiency is low and easy to make mistakes.

[0027] (2) Manual and semi-automatic equipment inventory, manual operation of semi-automatic equipment such as Personal Digital Assistant (PDA) or handheld terminal to scan barcodes or two-dimensional codes on goods, the system automatically enters the goods information, which ensures the accuracy of the goods information, but the efficiency is still very low.

[0028] (3) Automatic equipment inventory, manual setting of inventory tasks, automatic equipment such as unmanned aerial vehicles, autonomous mobile robots or fixed sensors deployed in a range, which can batch acquire goods information through scanning labels or machine vision and automatically input into the system, which saves the process of manual movement, realizes parallel processing of multiple inventory tasks and greatly improves the inventory efficiency, but the cost of equipment and environmental arrangement is higher.

[0029] Indoor positioning technology can be divided into three categories based on physical principles: acoustic-based, optical-based, and radio wave-based.

[0030] (1) Acoustic-based positioning can be further divided into ultrasonic wave (>20KHz) and audible sound (1-20KHz) positioning, which calculates distance through time of arrival algorithm, with high precision but slow speed (air speed is about 340m / s, much smaller than light speed), and sound speed is sensitive to environmental temperature and humidity, which needs to be strictly controlled, so the deployment cost is not low.

[0031] (2) Optical-based positioning can be further divided into infrared, visible light communication, and machine vision. Infrared and visible light communication calculate distance through time of arrival or angle of arrival algorithm, with high precision and strong anti-interference, but the non-shielded line-of-sight feature requires dense deployment of sensors, and both ends need to be able to actively emit signals, which is costly. Machine vision calculates the position of the camera through visual feature matching or image recognition positioning. Visual feature matching establishes the geometric relationship between images through the change of feature points or feature regions before and after movement. Image recognition positioning such as AprilTag establishes the geometric relationship between the camera and the image through the distortion of the image in the picture.

[0032] (3) Radio wave-based positioning can be subdivided into WIFI, Bluetooth, ZigBee, UWB, RFID, LoRa, millimeter wave radar, etc. The distance between the transceiver devices is calculated by algorithms such as time of arrival, angle of arrival, phase of arrival, and received signal strength. The essential difference lies in the frequency band used, signal bandwidth, signal modulation method, and device cooperation mechanism. In general, radio waves have poor high-frequency penetration, but high positioning resolution. The wider the signal bandwidth, the more accurate the measurable time or phase, and the higher the positioning accuracy. LoRa has low power consumption, long coverage distance, but the lowest accuracy, with a maximum of only meters. WIFI, Bluetooth, and ZigBee can achieve sub-meter positioning accuracy, but the frequency bands used are very crowded and are easily disturbed, requiring special signal modulation. UWB and millimeter wave radar have a large bandwidth and strong anti-interference, making them suitable for high-precision positioning, but the device power consumption and cost are high. In addition, the above technical means and active RFID require the devices at the positioning end to be powered on for a long time and can analyze the corresponding communication protocol, which has a high cost for large-scale deployment in warehouse goods positioning scenarios. Passive RFID generally uses passive tags or paper tags, which have low production costs and sub-meter or even centimeter-level positioning accuracy, making them ideal for large-scale, low-cost indoor positioning technology.

[0033] In view of the advantages and disadvantages of the above technologies, the present application uses passive RFID technology to locate goods, and uses passive RFID, machine vision, and inertial navigation to locate drones.

[0034] In one exemplary embodiment, as shown in Figure 1 and Figure 2 a warehouse goods high-precision positioning and automatic inventory method is provided, which is executed by a computer device and includes the following steps 101 to 104.

[0035] Step 101, constructing a fingerprint library. The fingerprint library includes the received signal strength of each reference tag. The reference tag is a passive radio frequency identification tag arranged on the shelves in the warehouse.

[0036] In one specific application example, according to the number and size of the shelves in the warehouse, passive RFID tags and virtual readers are arranged on the shelves in the warehouse. The received signal strength of each reference tag is collected by the virtual reader to construct a fingerprint library. The number of reference tags and virtual readers is multiple.

[0037] Specifically, based on the number and size of the shelves in the warehouse, spatial intervals are planned, and passive RFID tags are affixed at the intersections of the shelves as reference tags. The spatial distribution and coordinate parameters of the virtual readers are set according to the planned spatial intervals based on the number and size of the shelves. The virtual readers are distributed as evenly as possible on the horizontal plane, roughly covering the area of ​​all reference tags, and extending vertically to 2 to 3 layers. The drone sequentially passes through the designated virtual reader locations, recording the Received Signal Strength Indicator (RSSI) of the received reference tags, forming a fingerprint database. Each reference tag in the fingerprint database corresponds to a unique and distinct fingerprint, i.e., the detection results from multiple virtual readers.

[0038] Step 102: Generate scan points based on the inventory task and the layout of the shelves in the warehouse, and plan the flight path.

[0039] In a specific application example, the inventory task is either a location inventory task or a source inventory task. Users can choose to inventory by location or by source, selecting one or more locations / sources to be inventoried. After determining the task type and quantity, the user submits the inventory task.

[0040] For inventory location tasks, a specified number of scan points are automatically generated around a inventory location, and the overall flight path is planned by combining scan points from other inventory location tasks. The scan points are arranged horizontally around the inventory location at approximately 1 meter intervals (depending on shelf spacing and aisle width), with 5x5 points and 2 or 3 points vertically. Each inventory location is designed with scan points using the above method. If inventory locations are close together, overlapping scan points can be merged. Finally, the planned flight path passes through all scan points sequentially.

[0041] For inventory counting tasks, multiple paths and sparse scanning points are automatically planned across the entire shelving area. After the drone's initial flight obtains the approximate location of each item, the scanning points and flight path are replanned, and the drone takes a second flight to obtain the precise location of the items. The spacing of the sparse scanning points depends on the shelf spacing, aisle width, and the effective detection range of the RFID reader, aiming to cover the entire shelving area with as few scanning points as possible.

[0042] The flight path and scan points are uploaded to the drone, which then performs the corresponding flight mission.

[0043] Step 103: Using machine vision algorithms and RFID positioning methods, the drone is controlled to sequentially pass through the scanning points along the stated flight path. Simultaneously, the received signal strength of the tag under test is collected by the RFID reader carried by the drone. The tag under test is a passive RFID tag deployed on goods in the warehouse.

[0044] Conventional RFID readers are typically mounted on ceilings or walls. When a drone carries an RFID reader, it achieves the same suspended functionality as a conventional RFID reader through fixed-point scanning. The drone hovers at the scanning point for a period of time, and its RFID reader rotates 360° to scan the received signal strength of the tag under test.

[0045] Among these, machine vision algorithms are based on monocular feature tracking and inertial measurement units, such as the VINS-Mono algorithm or VINS-Fusion. RFID positioning methods can also utilize time of arrival, angle of arrival, and phase difference of arrival for multilateral positioning, or employ artificial neural networks and deep learning to match RFID fingerprints. In principle, the VINS-Mono algorithm can already achieve positioning, but inertial measurement is prone to accumulating errors. The original algorithm corrects the accumulated errors through machine vision feature tracking. This application introduces RFID-based positioning results as an additional reference to assist in error correction.

[0046] Step 104: Based on the received signal strength of the tag under test and the fingerprint database, use an iterative Bayes-based radio frequency identification fingerprint algorithm to calculate the spatial location of the tag under test in order to inventory the goods in the warehouse.

[0047] Specifically, the received signal strength of the tag under test is compared with the received signal strength of reference tags in the fingerprint database. The higher the similarity, the closer the tag under test is to the reference tag. Then, the position of the tag under test is determined based on the position of the reference tag whose similarity to the tag under test is greater than a set threshold.

[0048] The most common implementation of RFID fingerprinting algorithms is based on the Location Identification Based on Dynamic Active RFID Calibration (LANDMARC) algorithm, which is a mature existing algorithm. However, LANDMARC can only directly estimate the specific coordinates of the goods to be measured. This application combines an iterative Bayesian estimation algorithm with an RFID fingerprinting algorithm. In addition to the final calculated specific coordinates, it can also obtain the spatial distribution probability of the coordinates, and this spatial distribution probability is continuously updated as the number of scanning points increases during the scanning process.

[0049] In another exemplary embodiment, the warehouse cargo high-precision positioning and automatic inventory method further includes the following step 105.

[0050] Step 105: Determine and display the spatial probability distribution of goods based on the spatial location of the label to be tested. Specifically, the calculated spatial probability distribution of goods is visualized and displayed on the front end.

[0051] In another exemplary embodiment, the warehouse cargo high-precision positioning and automatic inventory method further includes the following step 106.

[0052] Step 106: Based on the spatial location of the tag under test, replan the scanning points and flight path, and control the UAV to pass through the scanning points sequentially according to the flight path. At the same time, collect the received signal strength of the tag under test through the radio frequency identification reader carried by the UAV, and then return to step 103.

[0053] Specifically, for the inventory task, the approximate location of each cargo is calculated based on the data obtained after the drone's first flight, and the scanning points and flight path are replanned to pinpoint the cargo location. After the drone's second flight passes all the set scanning points, it automatically returns to the hangar according to the planned path, collects the RFID tag information of all designated cargo, and completes the cargo location calculation, thus completing the inventory task.

[0054] For the inventory counting task, after passing all the set scanning points, the drone automatically returns to the hangar according to the planned path, collects RFID tag information around all the storage locations, completes the cargo location calculation, matches the cargo location coordinates with the storage location coordinate range, obtains a unique matching relationship between storage location and cargo, and completes the inventory counting task.

[0055] This application employs RFID technology for cargo positioning. RFID technology can simultaneously detect multiple tagged objects at a long distance. Combined with drones, it enables rapid and efficient large-scale cargo inventory checks, avoiding errors and human oversights. Furthermore, passive RFID tags are low-cost to manufacture, rewritable, and reusable, and are simple and convenient to deploy. A fusion solution of the VINS-Mono machine vision algorithm and RFID is used for drone positioning. The camera and inertial measurement unit required by VINS-Mono are standard equipment on most drones, and the only additional sensor required by the drone is a single RFID reader / writer. The entire solution is intelligent, efficient, and cost-effective, facilitating rapid deployment and widespread adoption.

[0056] In one exemplary embodiment, such as Figure 3 As shown, a warehouse cargo high-precision positioning and automatic inventory system is provided, including: fingerprint database construction module 301, path planning module 302, UAV control module 303 and position calculation module 304.

[0057] The fingerprint database construction module 301 is used to construct a fingerprint database. The fingerprint database includes the received signal strength of each reference tag. The reference tags are passive radio frequency identification (RFID) tags deployed on warehouse shelves.

[0058] The path planning module 302 is used to generate scan points based on the inventory task and the layout of the shelves in the warehouse, and to plan the flight path.

[0059] The UAV control module 303 uses machine vision algorithms and radio frequency identification (RFID) positioning methods to control the UAV 305 to sequentially pass through the scanning points along the flight path. During the flight of the UAV 305, the received signal strength of the tag under test is collected by the onboard RFID reader 306. The tag under test is a passive RFID tag 307 deployed on goods in the warehouse.

[0060] The location calculation module 304 is used to obtain the received signal strength of the tag to be tested collected by the UAV, and according to the received signal strength of the tag to be tested and the fingerprint database, it uses an iterative Bayes-based radio frequency identification fingerprint algorithm to calculate the spatial location of the tag to be tested in order to inventory the goods in the warehouse.

[0061] The fingerprint database construction module 301, path planning module 302, UAV control module 303 and position calculation module 304 mentioned above are all computer programs that execute on a computer device. Among them, the UAV control module 303 and the position calculation module 304 are connected to the external UAV 305 and interact with the UAV 305.

[0062] In addition, the system will update and visualize the cargo positioning process in real time, and display the distribution probability changes of the predicted cargo location synchronously at the front end. As the number of detection points passed by the UAV 305 increases, the predicted cargo distribution range gradually shrinks to the vicinity of the actual location.

[0063] In another exemplary embodiment, the warehouse cargo high-precision positioning and automatic inventory system also includes a drone 305, an RFID reader / writer 306, and passive RFID tags 307. That is, in addition to the computer program (fingerprint database construction module 301, path planning module 302, drone control module 303, and location calculation module 304) executed on the computer device, the warehouse cargo high-precision positioning and automatic inventory system can also include the drone 305 connected to the drone control module 303 and the location calculation module 304, the RFID reader / writer carried by the drone 305, the passive RFID tags 307 deployed on the warehouse shelves, and the passive RFID tags 307 deployed on the goods. The drone 305 transmits the received signal strength to the location calculation module 304 in real time via WIFI.

[0064] In one exemplary embodiment, a computer apparatus is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0065] In summary, compared with the prior art, the beneficial effects of this application include the following points.

[0066] (1) Improved positioning accuracy. The UAV positioning achieves sub-meter or even centimeter-level positioning accuracy through VINS-Mono machine vision technology and RFID positioning method, and acts as a fixed-point reader in the subsequent cargo positioning process. Cargo positioning achieves sub-meter-level positioning accuracy through RFID fingerprint algorithm based on iterative Bayes, which meets the accuracy requirements of general warehouse cargo positioning and achieves 100% accurate warehouse inventory positioning.

[0067] (2) Improved inventory efficiency. During inventory, the drone needs to scan a certain number of scanning points around the target location / source of goods, but it can scan multiple locations / sources of goods at the same time. Considering the number of locations / sources of goods within the scanning point array, as well as the hovering scanning time and the movement time between scans, the efficiency of drone inventory can reach an average of 30 items per minute under ideal conditions.

[0068] (3) The cost is low and easy to deploy. The positioning scheme adopts VINS-Mono machine vision and passive RFID technology. The required drone sensors are a camera, an inertial measurement unit, and an RFID reader. In addition, a number of passive RFID tags are required.

[0069] (4) Real-time visualization of cargo positioning process and iterative prediction results. The iterative prediction results adopt RFID fingerprinting method based on iterative Bayes estimation. As the number of observation points increases, the prediction result range shrinks and is displayed synchronously on the front end.

[0070] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0071] In this application, all actions to acquire signals, information, or data are carried out in compliance with the relevant data protection laws and policies of the country where the location is situated, and with the authorization granted by the owner of the relevant device.

[0072] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0073] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for high-precision positioning and automatic inventory counting of warehouse goods, characterized in that, The method includes: Based on the number and size of the shelves in the warehouse, spatial intervals are planned, and passive RFID tags are affixed at the intersections of the shelves as reference tags. Based on the number and size of the shelves, spatial intervals are planned, and the spatial distribution and coordinate parameters of virtual readers are set. The virtual readers are evenly distributed on the horizontal plane and cover the range of all reference tags. The drone passes through the set virtual reader positions in sequence, records the received signal strength of the reference tags, and forms a fingerprint database. The fingerprint database includes the received signal strength of each reference tag. Based on the inventory task and the layout of the warehouse shelves, scanning points are generated, and a flight path is planned. The inventory task can be a location inventory task or a source inventory task. For location inventory tasks, a specified number of scanning points are automatically generated around a location, and the overall flight path is planned by combining the scanning points from other location inventory tasks. The scanning points are centered on the location, with 5×5 points spaced 1 meter apart around it. Each location is designed with scanning points according to the above method. If the locations are close, some overlapping scanning points are merged. Finally, the flight path is planned to pass through all the scanning points in sequence. For source inventory tasks, multiple paths and sparse scanning points are automatically planned to traverse the entire shelf area. After the drone obtains the approximate location of each source through its first flight, the scanning points and flight path are replanned, and the drone takes a second flight to obtain the precise location of the source. The spacing of the sparse scanning points depends on the shelf spacing, aisle width, and the effective detection distance of the RFID reader. Using machine vision algorithms and radio frequency identification (RFID) positioning methods, the drone is controlled to pass through the scanning points sequentially along the flight path. Simultaneously, the received signal strength of the tag under test is collected by the RFID reader carried by the drone. The tag under test is a passive RFID tag deployed on goods in the warehouse. The machine vision algorithm is a VINS-Mono algorithm based on monocular feature tracking and inertial measurement unit. The RFID-based positioning results are introduced as an additional reference to help correct errors. Based on the received signal strength of the tag under test and the fingerprint database, an iterative Bayes-based radio frequency identification fingerprint algorithm is used to calculate the spatial location of the tag under test in order to inventory the goods in the warehouse. The scanning points and flight path are replanned based on the spatial location of the tag under test, and the drone is controlled to pass through the scanning points sequentially according to the flight path. At the same time, the received signal strength of the tag under test is collected by the radio frequency identification reader carried by the drone.

2. The method for high-precision positioning and automatic inventory counting of warehouse goods according to claim 1, characterized in that, There are multiple reference tags and multiple virtual readers / writers.

3. The warehouse cargo high-precision positioning and automatic inventory method according to claim 1, characterized in that, The method further includes: The spatial probability distribution of goods is determined and displayed based on the spatial location of the label to be tested.

4. A high-precision positioning and automatic inventory system for warehouse goods, characterized in that, The system is applied to the warehouse cargo high-precision positioning and automatic inventory method according to any one of claims 1-3, and the system comprises: A fingerprint database construction module is used to construct a fingerprint database; the fingerprint database includes the received signal strength of each reference tag; the reference tag is a passive radio frequency identification tag deployed on a warehouse shelf; The path planning module is used to generate scan points based on the inventory task and the layout of the shelves in the warehouse, and to plan the flight path. The drone control module is used to control the drone to pass through the scanning points sequentially according to the flight path using machine vision algorithms and radio frequency identification (RFID) positioning methods; during the drone's flight, the received signal strength of the tag under test is collected by the onboard RFID reader / writer; the tag under test is a passive RFID tag deployed on goods in the warehouse. The location calculation module is used to obtain the received signal strength of the tag to be tested collected by the drone, and to calculate the spatial location of the tag to be tested using an iterative Bayes-based radio frequency identification fingerprint algorithm based on the received signal strength of the tag to be tested and the fingerprint database, so as to carry out inventory of goods in the warehouse.

5. The warehouse cargo high-precision positioning and automatic inventory system according to claim 4, characterized in that, The system also includes: the drone, the RFID reader / writer, and the passive RFID tag.

6. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the warehouse goods high-precision positioning and automatic inventory method according to any one of claims 1-3.

Images