Intelligent warehouse checking method and system

Abstract

The invention discloses an intelligent warehouse checking method and system, and the method comprises the steps: combining a radio frequency identification unit with a visual detection unit through a radio frequency identification unit, the visual detection unit, a model building unit and a processing unit, obtaining the cargo stacking position information and the cargo stacking mode of a cargo pile, carrying out the modeling of the cargo pile through the model building unit, and carrying out the processing of the model building unit. The processing unit carries out information collection and comparison on the cargo models to realize cargo inventory, information collection of cargo inventory can be carried out through a combination mode of radio frequency identification and visual inspection, careful detection can be carried out according to the cargo stacking difference, and the cargo inventory efficiency is improved. The problem that in the prior art, a storage inventory system cannot timely identify an inventory error when coping with the inventory error is solved. Through a mode of combining the signal identification intensity and the three-dimensional spherical coordinates, the model is enabled to be combined with the three-dimensional rectangular coordinate system and the three-dimensional spherical coordinates at the same time, the single model construction element caused by using a single coordinate system is avoided, and the accuracy of model construction is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of warehouse inventory counting, and in particular to an intelligent warehouse inventory counting method and system. Background Art

[0002] Goods warehousing refers to the storage and management of goods in warehouses, which involves a series of links such as receiving, warehousing, storage, distribution and outbound delivery of goods. The purpose of warehousing is to provide reasonable, safe and effective storage conditions to ensure that the goods are not damaged during storage to meet the needs of production and sales.

[0003] Inventory, also known as commodity inventory, refers to the counting and verification of inventory goods. This process involves checking and counting the physical quantity and storage conditions of the existing goods in the warehouse in order to grasp the actual inventory situation and check with the book records to find out whether there are any differences in quantity, quality, storage conditions, etc. In short, inventory is a work of checking and counting the quantity and quality of stored goods, and it is also an important part of warehouse management. Inventory helps enterprises to accurately grasp the real-time situation of goods, and then carry out effective inventory management, including ordering, determination of inventory levels, and sales coordination. The inventory process also helps to discover problems in inventory management, such as lost, damaged or expired goods, so that timely measures can be taken to correct them.

[0004] Prior art CN 111661551 B discloses a warehouse inventory method and a warehouse system. The warehouse inventory method includes: issuing an inventory task to a robot; the inventory task includes the storage location of a target material box; receiving the recognition result reported by the robot; the recognition result includes the target material box information. Through the warehouse inventory method, it is possible to implement a full-box inventory of the target material box, quickly find the target material box with problems, and improve the efficiency of warehouse inventory. However, during the warehouse inventory process, since the inventory process is single and all of them are direct recognition results of the goods by the robot, it is impossible to correct the errors in the inventory in a timely manner when unexpected situations occur, which affects the quality of the inventory.

[0005] Therefore, it is necessary to provide a new warehouse inventory counting system and method to solve the above problems. Summary of the invention

[0006] The present invention overcomes the shortcomings of the prior art and provides an intelligent warehouse inventory counting method and system, aiming to solve the problem in the prior art that the warehouse inventory counting system cannot promptly identify errors when inventory counting occurs.

[0007] To achieve the above purpose, the technical solution adopted by the present invention is: an intelligent warehouse inventory system, comprising:

[0008] A radio frequency identification unit, used to read and record cargo data, the radio frequency identification unit comprising a radio frequency identification tag and a radio frequency identification inventory counting device, the radio frequency identification tag being fixedly arranged on the surface of the cargo, each cargo being provided with one radio frequency identification tag, the radio frequency identification inventory counting device being able to identify the radio frequency identification tag, and read and record the radio frequency identification data recorded on the radio frequency identification tag;

[0009] A visual inspection unit, used to perform image inspection on the surface of goods, wherein the visual inspection unit performs image recognition on the goods and pre-processes the image;

[0010] A model building unit, used to build a model for the collected RFID data and images, the model building unit is connected to the RFID inventory counting device and the visual inspection unit respectively, the model building unit reads the RFID data and images, obtains the stacking method and the stacking quantity of the goods, and builds a stacking model of the goods according to the stacking method and the stacking quantity of the goods;

[0011] A processing unit is used to record the stacking model and perform inventory counting of goods. The processing unit receives and records the stacking model data constructed by the model building unit, and controls the radio frequency identification unit and the visual inspection unit to work respectively. The processing unit is provided with goods reservation information, and the goods reservation information includes the quantity of goods.

[0012] In a preferred embodiment of the present invention, a moving mechanism and a height adjustment mechanism are respectively provided on the RFID inventory counting device, the moving mechanism drives the RFID inventory counting device to move horizontally, and the height adjustment mechanism adjusts the height of the RFID inventory counting device.

[0013] In a preferred embodiment of the present invention, the information identified on the RFID tag by the RFID inventory counting device includes RFID tag record information and RFID tag signal recognition strength.

[0014] In a preferred embodiment of the present invention, the detection views of the detection image of the goods by the visual inspection unit include a side view, a top view and a plurality of stereoscopic views.

[0015] In a preferred embodiment of the present invention, the model building unit extracts the shape features of the goods from the processing unit while reading the RFID data and the image, including but not limited to geometric shape, edge and corner features, and combines the shape features of the goods with the stacking method of the goods to establish a three-dimensional spherical coordinate model.

[0016] In a preferred embodiment of the present invention, in the spherical coordinate model of the model building unit, the parameters include radius, polar angle and azimuth, and the radius parameter, polar angle parameter and azimuth parameter are all calculated based on the spherical coordinate center.

[0017] To achieve the above-mentioned purpose, the second technical solution adopted by the present invention is: a method for using an intelligent warehouse inventory counting system, based on an intelligent warehouse inventory counting system, comprising the following steps:

[0018] S1: The RFID unit performs RFID scanning when the goods are put into storage and taken out of storage, and performs RFID scanning on the goods pile to record the stacking position of the goods, so as to obtain the stacking position information of the goods.

[0019] S2: The visual inspection unit collects and pre-processes images of the goods after storage to obtain the goods stacking location information.

[0020] S3: The model building unit combines and analyzes the cargo stacking position information collected by the radio frequency identification unit and the cargo stacking form information collected by the visual inspection unit to obtain a cargo stacking model.

[0021] S4: The processing unit compares the quantity of goods in the goods stacking model with the quantity of goods in the reserved goods information, and compares the difference between the quantity obtained by the inventory counting system and the quantity calculated by the warehouse entry and exit.

[0022] S5: The processing unit analyzes the differences in the inventory results, evaluates the accuracy of the inventory, and optimizes the radio frequency identification unit, the visual inspection unit and the model building unit to improve the accuracy.

[0023] In a preferred embodiment of the present invention, when comparing the difference between the quantity obtained by the inventory counting system and the quantity calculated by the warehouse entry and exit in S4, the inventory counting system is corrected according to the positive or negative difference.

[0024] In a preferred embodiment of the present invention, when performing inventory accuracy analysis in S5, analysis is performed from the aspects of comparative analysis, error rate analysis, repeatability test and blind test.

[0025] In a preferred embodiment of the present invention, the optimization of the radio frequency identification unit, the visual detection unit and the model building unit in S5 is performed from the aspects of redundant design, light compensation and iterative optimization algorithm respectively.

[0026] The present invention solves the defects existing in the background technology and has the following beneficial effects:

[0027] (1) The present invention provides an intelligent warehouse inventory method and system, which utilizes a radio frequency identification unit, a visual detection unit, a model building unit and a processing unit. The radio frequency identification unit and the visual detection unit are combined to obtain the cargo stacking position information and cargo stacking method of the cargo pile. The model building unit models the cargo pile, and the processing unit collects and compares information on the cargo model to implement cargo inventory. Compared with the warehouse inventory system in the prior art, the method can collect information on cargo inventory by combining radio frequency identification and visual detection, thereby improving the authenticity of information collection, improving the detection accuracy of cargo inventory, and performing detailed detection according to the differences in cargo stacking, thereby solving the problem that the warehouse inventory system in the prior art cannot timely identify errors in the inventory.

[0028] (2) In the present invention, the detection view of the visual detection unit includes a side view, a top view and several stereoscopic views, so that the visual detection unit can collect images of the cargo pile from multiple angles. Compared with the prior art, the image collection can be three-dimensional, thereby realizing the three-dimensional construction of the cargo pile image.

[0029] (3) In the present invention, when the model building unit builds the model, the model type is three-dimensional spherical coordinates, which can enable a plurality of goods to simultaneously build a cargo pile model from two aspects of angle and distance. Compared with the prior art, the model building parameters can be increased, making the model more detailed and ensuring the accuracy of model building.

[0030] (4) In the present invention, by combining the signal recognition strength and the three-dimensional spherical coordinates, the model combines the three-dimensional rectangular coordinate system and the three-dimensional spherical coordinates at the same time, so that the cargo pile can be modeled from the two coordinate systems. Compared with the prior art, the model construction method is optimized, the single model construction element caused by the use of a single coordinate system is avoided, and the accuracy of the model construction is ensured. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 It is a system structure diagram of a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0035] In the description of the present application, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the scope of protection of the present application. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the invention, unless otherwise specified, "multiple" means two or more.

[0036] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood by specific circumstances.

[0037] The intelligent warehouse inventory counting system in this application targets a pile of goods that is in the form of tightly stacked goods, that is, the stacking density in each layer of goods is consistent.

[0038] like Figure 1 As shown, an intelligent warehouse inventory system includes:

[0039] The RFID unit is used to read and record cargo data. The RFID unit includes an RFID tag and an RFID inventory counting device. The RFID tag is fixedly set on the surface of the cargo. Each cargo is provided with an RFID tag. The RFID inventory counting device can identify the RFID tag and read and record the RFID data recorded on the RFID tag.

[0040] An RFID tag is an electronic tag or transponder, which is a device that communicates with a reader / writer via radio waves to transmit data. It can be in the form of a card or a label. An RFID inventory device is a reader / writer or interrogator that sends and receives electromagnetic signals from an RFID tag and reads the data on the RFID tag.

[0041] The visual inspection unit is used to perform image inspection on the surface of the goods. The visual inspection unit recognizes the goods and pre-processes the images. The visual inspection unit includes an image detection component and an image processing system. The image detection component can be an image acquisition device such as a scanner and a camera, and the image processing system can be a processing system such as a computer. The pre-processing of the image is to denoise or enhance the effect of the acquired image.

[0042] The model building unit is used to build a model for the collected RFID data and images. The model building unit is connected to the RFID inventory counting device and the visual inspection unit respectively. The model building unit reads the RFID data and images to obtain the stacking method and stacking quantity of the goods, and builds the stacking model of the goods according to the stacking method and stacking quantity of the goods.

[0043] The processing unit is used to record the stacking model and perform the inventory of the goods. The processing unit receives and records the stacking model data constructed by the model building unit, and controls the radio frequency identification unit and the visual inspection unit to work respectively. The processing unit is provided with the goods reservation information, and the goods reservation information includes the quantity of the goods. The processing unit is a processing system such as a computer, and the processing unit is connected to the radio frequency identification unit, the visual inspection unit and the model building unit respectively.

[0044] The radio frequency identification unit, the visual detection unit, the model building unit and the processing unit are used to combine the radio frequency identification unit and the visual detection unit to obtain the cargo stacking position information and the cargo stacking method of the cargo pile. The model building unit models the cargo pile, and the processing unit collects and compares information on the cargo model to realize cargo inventory. The information collection of cargo inventory can be carried out by combining radio frequency identification and visual detection, which can improve the authenticity of information collection, improve the detection accuracy of cargo inventory, and can perform detailed detection according to the differences in cargo stacking, thereby solving the problem in the prior art that the warehouse inventory system cannot timely identify errors in inventory.

[0045] The RFID inventory counting device is provided with a moving mechanism and a height adjustment mechanism. The moving mechanism drives the RFID inventory counting device to move horizontally, and the height adjustment mechanism adjusts the height of the RFID inventory counting device. The moving mechanism drives the RFID inventory counting device to enter the cargo position, ensures that the RFID inventory counting device is correctly aligned with the cargo, and adjusts it to a suitable position so that the RFID inventory counting device can identify the RFID tag. The moving mechanism can be a component such as a guide rail that moves the RFID inventory counting device horizontally, and the height adjustment mechanism is a component such as a cylinder that adjusts the height of the RFID inventory counting device.

[0046] The information identified on the RFID tag by the RFID inventory counting device includes the RFID tag record information and the RFID tag signal recognition strength. The RFID tag record information is the specification data of the goods themselves when they enter the warehouse, including but not limited to the goods identification code, name, model and storage precautions; the RFID tag signal recognition strength is the signal strength received or sent by the RFID tag during wireless communication. The signal recognition strength of the RFID tag is affected by many factors, including the distance between the tag and the reader, possible interference sources in the environment, the power of the tag and the antenna design, etc. The main influencing factor is the distance between the reader and the electronic tag. The closer the tag is to the reader, the stronger the signal strength. After the RFID unit performs an RFID scan on the stacked goods, it can obtain the centroid position of the goods through the signal strength, the distance between the RFID tag and the RFID inventory counting device, and the goods specifications on the reserved information of the goods.

[0047] The distance between the RFID tag and the RFID counting device is determined by the path loss model, which is Among them, PL(d) is the signal strength (dBm) when the distance between the RFID counting device and the RFID tag is d, PL(d0) is the signal strength at the reference distance d0, n is the signal attenuation index, which ranges from 2 to 4, d0 is the reference distance, and N0 is a Gaussian random noise variable with a mean of 0 and a standard deviation of σ. The distance between the centroid of the cargo and the RFID tag can be known through the cargo specifications on the cargo reservation information, and the centroid position of the cargo can be obtained by combining the RFID counting device.

[0048] The detection views of the detection images of the goods by the visual detection unit include a side view, a top view and several stereoscopic views. The side view refers to the view on the cargo pile parallel to the side plane of the cargo pile, the top view is the view of the horizontal plane directly above the cargo pile, and the stereoscopic view refers to the view located above the cargo pile and can simultaneously record the top surface and two side surfaces of the cargo pile. The image detection component is fixedly set on the warehouse and the angle is adjusted to complete image collection from multiple angles, that is, multiple cameras are used to shoot from the side, top surface and upper side respectively. The detection views of the visual detection unit include a side view, a top view and several stereoscopic views, so that the visual detection unit can collect images of the cargo pile from multiple angles, can make the image collection three-dimensional, and realize the three-dimensional construction of the cargo pile image.

[0049] Through multiple angles of view, the appearance characteristics and surface information of the goods can be fully captured. The front view shows the front features of the goods, the side view reveals the height and side features of the goods, the top view provides information on the top of the goods and how the goods are arranged, and the stereo view combines this information to provide a three-dimensional perspective. Such all-round detection helps the system to more accurately identify and record cargo information and improves recognition accuracy.

[0050] While the model building unit reads the RFID data and images, it extracts the appearance features of the goods from the processing unit, including but not limited to geometric shapes, edge and corner features, and combines the appearance features of the goods with the way the goods are stacked to build a three-dimensional spherical coordinate model. The geometric shape, edge and corner features can make the packaging shape of the goods clear and know the specifications, shape and size of the goods. When analyzing the stacked image of the goods, the stacking method of the goods can be obtained, and after combining with the RFID data, the specific coordinates of each goods can be obtained. When the model building unit builds the model, the model type is three-dimensional spherical coordinates, which can enable several goods to simultaneously build a cargo stack model from two aspects of angle and distance, and can increase the model building parameters to make the model more detailed and ensure the accuracy of model building.

[0051] The convolutional neural network can obtain feature vectors from the side view, top view and stereoscopic view, and the cargo stacking method can be obtained through the multimodal deep learning model. The cargo stacking method is determined to be one of the stacking methods such as rectangular, pyramidal, cylindrical and truncated cone, and the side length of the cargo pile and the cargo specifications on the cargo reservation information are obtained according to the cargo stacking method, so as to obtain the quantity of cargo. The centroid of the internal cargo is predicted based on the quantity of cargo, the centroid of the outer cargo and the cargo specifications on the cargo reservation information.

[0052] Taking a rectangular cargo pile as an example, by confirming the length of each side of the cargo pile l x , l y , lz , calculate the overall volume of the cargo pile V = l x × y × z , where l z It is the height of the second top layer of the cargo pile. The volume of a single cargo is compared with the overall volume of the cargo pile, and the scattered cargo on the top layer of the cargo pile is added to get the quantity of cargo in the cargo pile.

[0053] In the spherical coordinate model of the model building unit, the parameters include radius, polar angle and azimuth, and the radius parameter, polar angle parameter and azimuth parameter are all calculated based on the spherical coordinate center. The midpoint of the spherical coordinate is the center of mass of the cargo pile, the radius is the distance between the midpoint of the spherical coordinate and the center of mass of the cargo, the polar angle is the angle between the projection of the line connecting the midpoint of the spherical coordinate and the center of mass of the cargo on the horizontal plane and the positive direction of the positive axis of the horizontal plane, and the azimuth is the angle between the line connecting the vertical axis to the point and the positive direction of the vertical axis. The positive axis of the horizontal plane can be the direction of cargo out of the warehouse, and the positive direction of the vertical axis is vertically upward. The model building unit adopts a neural network algorithm and a machine learning algorithm to analyze and fuse the collected data and the equipment geometric data to form the mapping linkage relationship, map the three-dimensional position information of the cargo to the actual space of the warehouse, and calculate the specific position of each cargo in the spherical coordinate system according to the position of the radio frequency identification inventory device and the viewing angle of the visual detection unit.

[0054] By combining the signal recognition strength and three-dimensional spherical coordinates, the model combines the three-dimensional rectangular coordinate system and the three-dimensional spherical coordinates at the same time, so that the cargo pile can be modeled from two coordinate systems, which optimizes the model construction method, avoids the single model construction element caused by the use of a single coordinate system, and ensures the accuracy of model construction. Through the signal recognition strength, the size calculation on the plane can be performed in the rectangular coordinate system, and the relative relationship between each cargo and the center of mass of the cargo pile can be calculated through the three-dimensional spherical coordinates. The combination of the two can ensure that the position of each cargo is more accurate.

[0055] In the process of establishing the three-dimensional spherical coordinates, the rectangular coordinates of several goods are converted into three-dimensional spherical coordinates, (x i ,y i ,z i ) is the rectangular coordinate of the centroid of each cargo based on the rectangular coordinate origin, and the rectangular coordinate origin is the center point of the ground area where each cargo is stacked, V i is the volume of each cargo, and the coordinates (X, Y, Z) of the spherical coordinate center of the three-dimensional spherical coordinate in the rectangular coordinate system have the relationship: When this coordinate is used as the spherical coordinate center of the three-dimensional spherical coordinate, the coordinates (r, θ, φ) of each item in the three-dimensional spherical coordinate are: Get the spherical coordinate model.

[0056] The number of samples in the cargo pile is reduced from the top, and a new spherical coordinate model of the reduced cargo pile is established. On the one hand, the new model is constructed based on the radio frequency identification unit and the visual recognition unit using a combination of radio frequency tags. However, the rectangular coordinates of the centroid of each cargo based on the rectangular coordinate origin are initially determined to be different from the center point of the ground area where each cargo is stacked. On the other hand, a mathematical method is used to redefine the spherical coordinate model after subtracting the reduced sample part from the spherical coordinate system. The two spherical coordinate models are compared, and any cargo is selected to find out whether the coordinates of the cargo in the two coordinate systems are consistent.

[0057] To achieve the above-mentioned purpose, the second technical solution adopted by the present invention is: a method for using an intelligent warehouse inventory counting system, based on an intelligent warehouse inventory counting system, comprising the following steps:

[0058] S1: The RFID unit performs RFID scanning when the goods are put into the warehouse and taken out of the warehouse respectively, and performs RFID scanning on the goods pile to record the stacking position of the goods, so as to obtain the stacking position information of the goods.

[0059] S2: The visual inspection unit collects and pre-processes images of the goods after they enter the warehouse to obtain information on the stacking form of the goods.

[0060] S3: The model building unit combines and analyzes the cargo stacking position information collected by the radio frequency identification unit and the cargo stacking form information collected by the visual inspection unit to obtain a cargo stacking model.

[0061] S4: The processing unit compares the quantity of goods in the goods stacking model with the quantity of goods in the reserved goods information, and compares the difference between the quantity obtained by the inventory counting system and the quantity calculated by the in and out warehouse.

[0062] S5: The processing unit analyzes the differences in the inventory results, evaluates the accuracy of the inventory, and optimizes the radio frequency identification unit, visual inspection unit, and model building unit to improve accuracy.

[0063] In S4, when comparing the difference between the quantity obtained by the inventory counting system and the quantity calculated by the in and out warehouse, the inventory counting system is corrected according to the positive or negative difference. If the difference is positive, that is, the inventory count is more than the expected number, check whether the goods are counted repeatedly or whether there are unrecorded warehouse operations; if the difference is negative, that is, the inventory count is less than the expected number, check whether there are any goods lost, damaged or omitted during the inventory. When the processing unit finds that the difference is not zero, the RFID tag data of the goods recorded in the warehouse by the processing unit is compared with the RFID tag of the goods scanned by the RFID unit. Compare which goods have lost inventory information.

[0064] In S5, when analyzing the differences in the inventory results, the goods are classified and trend analyzed. Difference classification refers to the summary of the causes of inventory differences. After sorting, the causes such as wrong inbound and outbound counting, damaged labels, and misplaced goods are summarized and analyzed to determine the proportion of each cause in the inventory differences. Trend analysis refers to trend analysis of historical inventory data to obtain the trend of the proportion of each cause of inventory differences in the warehouse. The processing center makes targeted adjustments based on the difference classification and trend analysis trends.

[0065] When performing inventory accuracy analysis in S5, the analysis is conducted from the aspects of comparative analysis, error rate analysis, repeatability test and blind test. Comparative analysis can evaluate the accuracy of the system by comparing the data obtained by the inventory system with the actual inventory data. If the data obtained by the system is highly consistent with the actual inventory, it means that the accuracy of the inventory system is high; error rate analysis calculates the difference ratio between the system inventory results and the actual inventory to obtain the error rate, that is, the ratio of the number of inventory counts to the number of inventory differences. The lower the error rate, the higher the accuracy of the system; repeatability test is to count the same batch of goods multiple times and analyze the consistency of the results. If the results of each inventory count are highly consistent, it means that the accuracy and stability of the system are good; blind test method is to conduct inventory test without telling the system the specific goods information. Through blind test, the system's recognition and inventory capabilities in unknown situations can be tested, thereby evaluating its accuracy.

[0066] In S5, the optimization of the RFID unit, visual inspection unit and model building unit is carried out from the aspects of redundant design, light compensation and iterative optimization algorithm. Redundant design refers to the deployment of multiple RFID inventory counting devices in areas with large inventory difference rates to achieve cross-validation of data and reduce single point failures and recognition errors; light compensation refers to the increase of image quality by adjusting the light brightness near the cargo pile to improve the image exposure of the visual inspection unit; the iterative optimization algorithm improves the prediction accuracy of the model by continuously adjusting and optimizing the model parameters. The model building unit references historical data for preliminary model building and optimizes the model in multiple model building.

[0067] The above is based on the ideal embodiment of the present invention. Through the above description, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention. The technical scope of the present invention is not limited to the content in the specification, and the technical scope must be determined according to the scope of the claims.

Claims

1. An intelligent warehouse inventory system, characterized in that: include: A radio frequency identification unit, used to read and record cargo data, the radio frequency identification unit comprising a radio frequency identification tag and a radio frequency identification inventory counting device, the radio frequency identification tag being fixedly arranged on the surface of the cargo, each cargo being provided with one radio frequency identification tag, the radio frequency identification inventory counting device being able to identify the radio frequency identification tag, and read and record the radio frequency identification data recorded on the radio frequency identification tag; A visual inspection unit, used to perform image inspection on the surface of goods, wherein the visual inspection unit performs image recognition on the goods and pre-processes the image; A model building unit, used to build a model for the collected RFID data and images, the model building unit is connected to the RFID inventory counting device and the visual inspection unit respectively, the model building unit reads the RFID data and images, obtains the stacking method and the stacking quantity of the goods, and builds a stacking model of the goods according to the stacking method and the stacking quantity of the goods; A processing unit is used to record the stacking model and perform inventory counting of goods. The processing unit receives and records the stacking model data constructed by the model building unit, and controls the radio frequency identification unit and the visual inspection unit to work respectively. The processing unit is provided with goods reservation information, and the goods reservation information includes the quantity of goods.

2. The intelligent warehouse inventory counting system according to claim 1, characterized in that: The RFID inventory counting device is provided with a moving mechanism and a height adjustment mechanism respectively. The moving mechanism drives the RFID inventory counting device to move horizontally, and the height adjustment mechanism adjusts the height of the RFID inventory counting device.

3. The intelligent warehouse inventory counting system according to claim 1, characterized in that: The information identified on the RFID tag by the RFID inventory counting device includes RFID tag record information and RFID tag signal identification strength.

4. The intelligent warehouse inventory counting system according to claim 1, characterized in that: The detection views of the detection image of the goods by the visual inspection unit include a side view, a top view and a plurality of stereoscopic views.

5. The intelligent warehouse inventory counting system according to claim 1, characterized in that: The model building unit extracts the shape features of the goods from the processing unit while reading the RFID data and the image, including but not limited to geometric shape, edge and corner features, and combines the shape features of the goods with the stacking method of the goods to build a three-dimensional spherical coordinate model.

6. The intelligent warehouse inventory counting system according to claim 5, characterized in that: In the spherical coordinate model of the model building unit, the parameters include radius, polar angle and azimuth, and the radius parameter, polar angle parameter and azimuth parameter are all calculated based on the spherical coordinate center.

7. A method for using an intelligent warehouse inventory counting system, based on an intelligent warehouse inventory counting system according to any one of claims 1 to 6, characterized in that: The following steps are involved: S1: The RFID unit performs RFID scanning when the goods are put into storage and taken out of storage, and performs RFID scanning on the goods pile to record the stacking position of the goods, so as to obtain the stacking position information of the goods; S2: The visual inspection unit collects and preprocesses images of the goods after storage to obtain the goods stacking location information; S3: the model building unit combines and analyzes the cargo stacking position information collected by the radio frequency identification unit and the cargo stacking form information collected by the visual inspection unit to obtain a cargo stacking model; S4: The processing unit compares the quantity of goods in the goods stacking model with the quantity of goods in the reserved goods information, and compares the difference between the quantity obtained by the inventory counting system and the quantity calculated by the in and out warehouse; S5: The processing unit analyzes the differences in the inventory results, evaluates the accuracy of the inventory, and optimizes the radio frequency identification unit, the visual inspection unit and the model building unit to improve the accuracy.

8. The method for using the intelligent warehouse inventory counting system according to claim 7, characterized in that: In the above S4, when comparing the difference between the quantity obtained by the inventory counting system and the quantity calculated by the warehouse entry and exit, the inventory counting system is corrected according to the positive or negative difference.

9. The method for using the intelligent warehouse inventory counting system according to claim 7, characterized in that: When performing inventory accuracy analysis in S5, analysis is performed from the aspects of comparative analysis, error rate analysis, repeatability test and blind test.

10. The method for using the intelligent warehouse inventory counting system according to claim 7, characterized in that: In S5, the optimization of the radio frequency identification unit, the visual detection unit and the model building unit is performed from the aspects of redundant design, light compensation and iterative optimization algorithm.

Images