Wire storehouse storage intelligent identification system and method

Abstract

The invention discloses a wire warehouse storage intelligent identification system and method. The method comprises the following steps: S1, collecting a multi-angle image and identifying wire features; s2, mapping an identification result to a warehouse space coordinate to generate a warehouse state diagram; s3, reading the electronic tag and binding the electronic tag with the identification target to confirm the identity; s4, constructing a dynamic inventory database and updating the state of the wire rod in real time; s5, receiving a warehouse-out request, planning a path and generating a carrying instruction; s6, controlling the automatic equipment to execute carrying and monitoring path deviation; and S7, after carrying is completed, the inventory state is updated, and an ex-warehouse log is recorded. According to the invention, full-process automatic identification and accurate warehouse-in and warehouse-out control of wire storage are realized, and the storage efficiency and the management intelligence level are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent warehousing and logistics automation technology, and in particular to an intelligent identification system and method for wire warehouse storage. Background Art

[0002] Against the backdrop of the rapid development of intelligent manufacturing and warehouse management, the demand for wire materials in industries such as electricity, communications, and construction continues to grow. Wire products are typically large in size, diverse in variety, densely stacked, and irregularly coiled, placing higher demands on accuracy and efficiency in their warehouse management. Traditional wire warehouse management methods mostly rely on manual visual identification and registration, which not only presents problems such as delayed information entry, difficulty in locating materials, and low efficiency in warehousing and outbound operations, but also easily leads to management risks such as misjudgment, wrong shipments, or inventory discrepancies in large-scale, multi-batch operations. In addition, manual identification often makes mistakes when dealing with situations such as unclear stacking status or label detachment, which seriously restricts the refinement and intelligent development of warehousing systems.

[0003] In recent years, some companies have begun experimenting with introducing image recognition and radio frequency identification technologies for assisted warehouse management. These systems capture on-site images through cameras, enabling automatic detection of certain objects, while simultaneously using radio frequency devices to read electronic tags to improve recognition efficiency. However, these systems often struggle with data fusion and precise matching. Image recognition is susceptible to occlusion, lighting, or viewing angle limitations, resulting in unstable results. Radio frequency identification struggles to provide the specific spatial location and stacking relationships of wires, making it impossible to accurately locate and identify wires using a single technology. Furthermore, most current image recognition systems only implement target detection and lack the ability to map to physical coordinate space, making them incapable of supporting subsequent operations such as path planning and scheduling optimization.

[0004] During the outbound delivery process, traditional systems typically design transport routes based on static routing or pre-set rules, making them incapable of coping with the complex and changing stocking density and dynamic obstacles found in warehouses. The lack of real-time spatial perception and path replanning mechanisms results in low outbound delivery efficiency and high transport failure rates. Furthermore, most systems lack a unified dynamic inventory database, real-time updates on inventory status, and anomaly detection mechanisms, preventing them from achieving closed-loop information management throughout the entire process.

[0005] Therefore, existing wire warehouse management technologies still have significant deficiencies in multi-source information fusion, spatial mapping accuracy, dynamic path planning, handling execution feedback, and inventory update consistency. There is an urgent need for an intelligent warehouse identification method that integrates visual recognition and radio frequency identification, and has real-time spatial perception and task closed-loop control capabilities, to achieve full-process automation and data traceability in wire warehouses, from identification, confirmation, warehouse building, scheduling, to execution. This invention proposes a technical solution to address the above-mentioned problems, effectively resolving key bottlenecks in existing technologies, such as low recognition accuracy, unclear location, strong scheduling rigidity, and untimely data updates.

[0006] Therefore, how to provide an intelligent identification system and method for wire warehouse storage is an urgent problem that technicians in this field need to solve. Summary of the Invention

[0007] One purpose of the present invention is to propose an intelligent identification method for wire warehouse storage based on the fusion of image recognition and radio frequency identification. The present invention makes full use of multi-angle image acquisition technology, deep convolutional neural network target detection model and RFID identity recognition technology, and describes in detail the full-process automated operation from wire image recognition, position mapping, identity confirmation to outbound path planning and dynamic handling control. It has the advantages of high recognition accuracy, accurate positioning, efficient scheduling and strong real-time inventory management.

[0008] According to an embodiment of the present invention, a method for intelligently identifying wire warehouse storage includes the following steps: S1. Collect multi-angle image data in the warehouse, use convolutional neural networks to detect wire targets, extract categories, sizes and stacking locations, and generate recognition results; S2. Integrate the recognition results with the warehouse structure diagram, perform spatial mapping, and generate warehouse status data including wire category, location, and hierarchy. S3. Use radio frequency equipment to read the wire electronic tag, obtain batch and specification information, and associate and confirm it with the warehouse status data; S4. Build a dynamic inventory database to record the type, specification, location and status of wires, and continuously update it during the operation process; S5. Receive outbound task requests, analyze target requirements, plan the optimal route based on the inventory database and warehouse status data, and generate transportation instructions; S6. Control the automatic handling equipment to perform the handling operation, and use the image recognition module to monitor in real time. If a deviation is identified, the path replanning is triggered; S7. After the handling is completed, the inventory database is updated, the outbound status and operation log are recorded, and closed-loop management of the entire process is achieved.

[0009] Optionally, S1 includes: collecting image data through several angle cameras arranged in the warehouse, analyzing the images using several pre-trained scale convolutional neural network models, identifying the category, appearance size, position coordinates in the image and stacking level information of each wire, and forming a feature data set of the wire target.

[0010] Optionally, the S2 specifically includes: S21. Based on the obtained target features of each wire rod, extract the corresponding image position coordinates and stacking level information as input data required for spatial positioning; S22. Calling a two-dimensional spatial layout diagram corresponding to the actual structure of the wire warehouse, preliminarily projecting the wire position coordinates in the image to corresponding physical area position points in the warehouse layout diagram; S23. Perform coordinate conversion on the preliminary projection results based on the camera installation parameters, shooting angle, and viewing angle conversion model, completing the transformation operation from the image coordinates to the actual physical coordinates of the warehouse, and generating positioning coordinates consistent with the physical layout; S24. Combining the stacking level information of each wire target with the identified size characteristics, estimating the actual height in the warehouse space, and determining the vertical placement order and overall occupied space range of the wires in the current area; S25. For all wire targets that have completed coordinate transformation and height estimation, spatial position integration is performed to construct a three-dimensional spatial mapping data set, and spatial overlap conflicts and boundary rationality are checked and processed; S26. The processed three-dimensional spatial mapping data is converted into a structured storage status diagram, which includes the category, size, physical location, and stacking level information of each wire.

[0011] Optionally, the S3 specifically includes: S31. Setting up a radio frequency identification reader in the storage environment to scan the wire material entity at the target position for image recognition and extracting the unique identification code, batch number, and specification model from the corresponding electronic tag; S32. Match the identification data extracted from each electronic tag with the wire targets located in the warehouse status structure diagram, perform fusion comparison based on the position coordinates, size information, and category attributes, and generate a matching relationship data table; S33. Establish a one-to-one binding relationship for the items in the generated matching relationship data table whose confidence level is higher than the set threshold, complete the identity confirmation of the image recognition target and the label information, and write the confirmation result into the corresponding wire material entry in the warehouse status structure diagram; S34. For the wire target whose identity has been confirmed, the number field and the specification field in the structure diagram entry are updated to construct a fusion feature set with spatial location, category attributes and identity identification; S35. For wire targets that are not matched successfully or have conflicts, they are uniformly marked as pending review status and output as abnormal data records.

[0012] Optionally, the fusion matching operation of step S32 includes: calculating the matching confidence between the image recognition target and the radio frequency tag data based on the spatial position coordinates, size characteristics and category attributes of each image recognition target; the confidence is weightedly scored based on the three dimensions of spatial proximity, size similarity and category consistency, and the matching result with the highest confidence is bound as the only corresponding relationship.

[0013] Optionally, the S4 specifically includes: S41, extracting the category, specification, unique identification code, physical location, and stacking level information of each wire target based on the fused feature set, and constructing a wire inventory data entry; S42. All inventory data items are uniformly entered into a dynamic inventory database, and a standardized data structure is established for task invocation. Each item record includes the wire number, specification model, category, location coordinates, and stacking level; S43. During the execution of the warehousing operation, status feedback information from the image recognition module is received in real time, and the location coordinates and stacking level fields of the relevant entries in the inventory database are dynamically updated; S44. Perform integrity and consistency checks on the updated inventory data. If any entries with duplicate locations, missing status, or abnormal stacking relationships are found, they are recorded as abnormal data and output to the scheduling system for troubleshooting reference. S45. Output the dynamically updated inventory database to the inbound and outbound scheduling module for use in transportation path planning, task allocation, and resource scheduling steps, thereby forming a real-time and available wire inventory information basis.

[0014] Optionally, the S5 specifically includes: S51, receiving a delivery task request, extracting the wire material category, specification, and target quantity information contained in the task as delivery demand input; S52. Retrieve wire records that meet the requirements of the outbound task based on the in-stock items in the inventory database, sort them according to their location coordinates and stacking level priorities, and generate a list of candidate outbound wires. S53. Build a path model based on the warehouse structure diagram and the physical location of the candidate wires, perform path planning operations, and determine the shortest path from the current location of the wires to the delivery point, excluding obstacle areas and high-density stacking areas. S54. After completing the path planning, the feasibility of each path is evaluated and sorted in combination with the execution parameters of the handling equipment to determine the order of outbound delivery and the corresponding handling target location; S55: Generate a handling instruction set, which includes a warehouse-out sequence, a unique number of the corresponding wire material, spatial coordinates, and target position coordinates, and output the instruction set to the automatic handling module.

[0015] Optionally, the path planning of step S53 includes: constructing a dynamic obstacle avoidance graph structure based on the obstacle areas and stacking density information marked in the warehouse structure diagram; performing a path search operation on the graph structure, giving priority to path nodes with lower traffic density, excluding all obstacle areas marked as inaccessible, and obtaining the shortest effective path between the current position of the wire and the delivery point.

[0016] Optionally, the S6 specifically includes: controlling the automatic handling equipment to perform outbound operations one by one according to the handling instruction set, and collecting the current position of the wire and the deviation of the handling path in real time through the image recognition module during the handling process. If a position deviation or path blockage is detected, the path replanning mechanism is immediately triggered to regenerate a feasible path and update the handling instructions.

[0017] According to an embodiment of the present invention, a wire warehouse intelligent storage identification system includes the following modules: Image acquisition module, used to obtain multi-angle images of the warehouse and provide image input source; Target recognition module, used to identify wire targets in the image and extract their feature data; Spatial mapping module, used to convert the wire position in the image into physical coordinates of the warehouse; Tag reading module, used to obtain the RFID information of the wire and extract the number and specifications; Information fusion module, used to match and bind radio frequency data with image targets; Inventory management module, used to build and update a dynamic inventory database and store structured information; Path planning module, used to generate the optimal transportation path based on inventory and structure diagrams; The instruction generation module is used to output the transport execution instruction including the number and coordinates; The transport control module is used to drive the automatic equipment to perform transport and monitor the task status; The anomaly detection module is used to identify path deviations and data conflicts and trigger replanning.

[0018] The beneficial effects of the present invention are: (1) This invention achieves accurate identification and three-dimensional position mapping of wire entities by integrating multi-angle image acquisition with a deep neural network recognition algorithm, combining the spatial characteristics and stacking level information of wires in images. It also uses radio frequency identification technology to verify and bind wire identities, overcoming the problems of single visual recognition being susceptible to occlusion and radio frequency positioning being inaccurate, significantly improving the accuracy and consistency of wire inventory status recognition.

[0019] (2) This invention constructs a spatial path model based on a physical layout diagram and inventory data. Combining outbound task requirements with real-time inventory status, it uses an optimization algorithm to dynamically plan paths and arrange handling sequences, generating a structured handling instruction set. During the handling process, the device path deviation is monitored through image recognition, and the system has the ability to self-repair and replan paths, enabling efficient and stable execution of handling operations.

[0020] (3) The present invention establishes a unified dynamic inventory database model, which synchronizes the wire identification results, location information and in-and-out status in real time, supports the identification and elimination of abnormal entries, and provides data support for in-and-out task scheduling. This mechanism effectively solves the problems of data update lag, information islands and status inconsistency in traditional warehousing, and realizes the data closed loop and intelligent management of the entire process of wire warehousing. BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is an overall flow chart of an intelligent identification method for wire warehouse storage proposed by the present invention; Figure 2 This is a flowchart of the mapping of image coordinates and physical locations of a wire warehouse intelligent identification method proposed by the present invention; Figure 3 This is a flowchart of the fusion of radio frequency data and image recognition targets for the intelligent identification method for wire warehouse storage proposed by the present invention. DETAILED DESCRIPTION The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0021] refer to Figure 1-3 , a wire warehouse storage intelligent identification system and method, comprising the following steps: S1, collecting multi-angle image data in the warehouse, using a convolutional neural network to detect wire targets, extracting categories, sizes and stacking positions, and generating identification results; S2. Integrate the recognition results with the warehouse structure diagram, perform spatial mapping, and generate warehouse status data including wire category, location, and hierarchy. S3. Use radio frequency equipment to read the wire electronic tag, obtain batch and specification information, and associate and confirm it with the warehouse status data; S4. Build a dynamic inventory database to record the type, specification, location and status of wires, and continuously update it during the operation process; S5. Receive outbound task requests, analyze target requirements, plan the optimal route based on the inventory database and warehouse status data, and generate transportation instructions; S6. Control the automatic handling equipment to perform the handling operation, and use the image recognition module to monitor in real time. If a deviation is identified, the path replanning is triggered; S7. After the handling is completed, the inventory database is updated, the outbound status and operation log are recorded, and closed-loop management of the entire process is achieved.

[0022] The present invention solves the problems of inconsistent information, chaotic entry and exit, and difficult positioning in traditional wire warehousing by constructing a full-process intelligent identification method for wire warehousing that integrates image recognition, spatial mapping, radio frequency identification, inventory updating, and automatic scheduling. It realizes closed-loop management from identification to execution, and improves identification accuracy and outbound efficiency.

[0023] In this embodiment, S1 includes: collecting image data through several angle cameras arranged in the warehouse, analyzing the images using several pre-trained scale convolutional neural network models, identifying the category, appearance size, position coordinates in the image and stacking level information of each wire, and forming a feature data set of the wire target.

[0024] For the image acquisition and recognition links, the present invention uses multi-angle camera and deep neural network model to extract wire position, size and stacking level, achieving high-precision extraction of wire information in complex stacking scenarios, significantly improving front-end recognition quality and data integrity.

[0025] In this embodiment, S2 specifically includes: S21. Based on the obtained target features of each wire rod, extract the corresponding image position coordinates and stacking level information as input data required for spatial positioning; S22. Calling a two-dimensional spatial layout diagram corresponding to the actual structure of the wire warehouse, preliminarily projecting the wire position coordinates in the image to corresponding physical area position points in the warehouse layout diagram; S23. Perform coordinate conversion on the preliminary projection results based on the camera installation parameters, shooting angle, and viewing angle conversion model, completing the transformation operation from the image coordinates to the actual physical coordinates of the warehouse, and generating positioning coordinates consistent with the physical layout; S24. Combining the stacking level information of each wire target with the identified size characteristics, estimating the actual height in the warehouse space, and determining the vertical placement order and overall occupied space range of the wires in the current area; S25. For all wire targets that have completed coordinate transformation and height estimation, spatial position integration is performed to construct a three-dimensional spatial mapping data set, and spatial overlap conflicts and boundary rationality are checked and processed; S26. The processed three-dimensional spatial mapping data is converted into a structured storage status diagram, which includes the category, size, physical location, and stacking level information of each wire.

[0026] The present invention introduces a mechanism for converting image coordinates into physical warehouse coordinates during the spatial mapping process, and generates three-dimensional position information in combination with the stacking height, thereby achieving precise positioning of wires in physical space, supporting subsequent path planning and inventory modeling operations, and breaking through the limitation of "inaccurate two-dimensional positioning" of traditional image recognition.

[0027] In this embodiment, S3 specifically includes: S31. Setting up a radio frequency identification reader in the storage environment to scan the wire material entity at the target position for image recognition and extracting the unique identification code, batch number, and specification model from the corresponding electronic tag; S32. Match the identification data extracted from each electronic tag with the wire targets located in the warehouse status structure diagram, perform fusion comparison based on the position coordinates, size information, and category attributes, and generate a matching relationship data table; S33. Establish a one-to-one binding relationship for the items in the generated matching relationship data table whose confidence level is higher than the set threshold, complete the identity confirmation of the image recognition target and the label information, and write the confirmation result into the corresponding wire material entry in the warehouse status structure diagram; S34. For the wire target whose identity has been confirmed, the number field and the specification field in the structure diagram entry are updated to construct a fusion feature set with spatial location, category attributes and identity identification; S35. For wire targets that are not matched successfully or have conflicts, they are uniformly marked as pending review status and output as abnormal data records.

[0028] The present invention establishes a unique binding relationship and performs abnormal marking by fusing radio frequency tags with image recognition results, effectively solving problems such as label loss, recognition errors and target mismatch in traditional warehousing, and improving the stability and automation level of identity confirmation.

[0029] In this embodiment, the fusion matching operation of step S32 includes: calculating the matching confidence between the image recognition target and the radio frequency tag data based on the spatial position coordinates, size characteristics and category attributes of each image recognition target; the confidence is weighted based on the three dimensions of spatial proximity, size similarity and category consistency, and the matching result with the highest confidence is bound as the only corresponding relationship.

[0030] The present invention introduces a multi-dimensional confidence calculation mechanism in the fusion matching process to achieve comprehensive comparison of position, size and category, greatly improving the accuracy of tag matching and avoiding the problem of misbinding caused by visual misrecognition and RFID reading interference.

[0031] In this embodiment, the S4 specifically includes: S41, extracting the category, specification, unique identification code, physical location, and stacking level information of each wire target based on the fused feature set, and constructing a wire inventory data entry; S42. All inventory data items are uniformly entered into a dynamic inventory database, and a standardized data structure is established for task invocation. Each item record includes the wire number, specification model, category, location coordinates, and stacking level; S43. During the execution of the warehousing operation, status feedback information from the image recognition module is received in real time, and the location coordinates and stacking level fields of the relevant entries in the inventory database are dynamically updated; S44. Perform integrity and consistency checks on the updated inventory data. If any entries with duplicate locations, missing status, or abnormal stacking relationships are found, they are recorded as abnormal data and output to the scheduling system for troubleshooting reference. S45. Output the dynamically updated inventory database to the inbound and outbound scheduling module for use in transportation path planning, task allocation, and resource scheduling steps, thereby forming a real-time and available wire inventory information basis.

[0032] The present invention constructs a dynamic inventory database system that can update the spatial status and inventory information of wires in real time, supports anomaly detection and data consistency verification, significantly improves the accuracy of inventory information and system stability, and solves the data lag problem of existing systems.

[0033] In this embodiment, the S5 specifically includes: S51, receiving a delivery task request, extracting the wire material category, specification, and target quantity information contained in the task as delivery demand input; S52. Retrieve wire records that meet the requirements of the outbound task based on the in-stock items in the inventory database, sort them according to their location coordinates and stacking level priorities, and generate a list of candidate outbound wires. S53. Build a path model based on the warehouse structure diagram and the physical location of the candidate wires, perform path planning operations, and determine the shortest path from the current location of the wires to the delivery point, excluding obstacle areas and high-density stacking areas. S54. After completing the path planning, the feasibility of each path is evaluated and sorted in combination with the execution parameters of the handling equipment to determine the order of outbound delivery and the corresponding handling target location; S55: Generate a handling instruction set, which includes a warehouse-out sequence, a unique number of the corresponding wire material, spatial coordinates, and target position coordinates, and output the instruction set to the automatic handling module.

[0034] The present invention combines outbound tasks with current inventory status, performs dynamic path planning and outbound priority sorting, and generates standardized handling instructions, thereby achieving efficient scheduling and task execution in complex stacking environments, significantly improving outbound accuracy and operational efficiency.

[0035] In this embodiment, the path planning of step S53 includes: constructing a dynamic obstacle avoidance graph structure based on the obstacle areas and stacking density information marked in the warehouse structure diagram; performing a path search operation on the graph structure, giving priority to path nodes with lower traffic density, excluding all obstacle areas marked as inaccessible, and obtaining the shortest effective path between the current position of the wire and the outbound point.

[0036] The present invention introduces an obstacle avoidance and density-aware path selection mechanism, establishes a dynamic obstacle avoidance graph structure, effectively avoids high-density areas and path congestion points, improves the flexibility and safety of path planning, and adapts to complex and changeable warehousing operation environments.

[0037] In this embodiment, S6 specifically includes: controlling the automatic handling equipment to perform outbound operations one by one according to the handling instruction set, and collecting the current position of the wire and the deviation of the handling path in real time through the image recognition module during the handling process. If a position deviation or path blockage is detected, the path replanning mechanism is immediately triggered to regenerate a feasible path and update the handling instructions.

[0038] During the transport process, the present invention uses image recognition to feedback the transport status, realizes real-time monitoring and response processing of path deviations and anomalies, ensures task continuity and safety, and improves the intelligence level and operational stability of the unmanned transport system.

[0039] According to an embodiment of the present invention, a wire warehouse intelligent storage identification system includes the following modules: Image acquisition module, used to obtain multi-angle images of the warehouse and provide image input source; Target recognition module, used to identify wire targets in the image and extract their feature data; Spatial mapping module, used to convert the wire position in the image into physical coordinates of the warehouse; Tag reading module, used to obtain the RFID information of the wire and extract the number and specifications; Information fusion module, used to match and bind radio frequency data with image targets; Inventory management module, used to build and update a dynamic inventory database and store structured information; Path planning module, used to generate the optimal transportation path based on inventory and structure diagrams; The instruction generation module is used to output the transport execution instruction including the number and coordinates; The transport control module is used to drive the automatic equipment to perform transport and monitor the task status; The anomaly detection module is used to identify path deviations and data conflicts and trigger replanning.

[0040] The system proposed in the present invention has a clear modular structure, a clear functional division of labor, high scalability and integrability, and a closed-loop data flow connection between modules, forming an intelligent identification and scheduling system suitable for wire storage environments, which effectively realizes automatic identification, precise positioning and efficient inbound and outbound control.

[0041] Example 1: To verify the feasibility of this invention, we applied it to a large-scale industrial wire turnover warehouse scenario, characterized by a wide variety of wires, high stockpiling density, and difficulty in accurate manual management. The site, covering over 4,500 square meters, features eight main storage areas, with over 200 rolls entering and leaving the warehouse daily. Traditional management relies on manual identification and record-keeping, resulting in low recognition efficiency, high error rates, and inconsistent inventory data.

[0042] This project deployed an intelligent wire storage identification system based on the fusion of image recognition and radio frequency identification, as proposed in this paper. The system utilizes 18 high-definition cameras installed throughout the warehouse's main aisles and stacking locations to provide multi-angle image coverage. Using a deep convolutional neural network model, the system detects and extracts features from wires within the images. The system automatically identifies wire types (such as copper cables, communications cables, and steel strands), their dimensions (diameters ranging from 18mm to 50mm), their spatial coordinates, and their stacking level. Simultaneously deployed RFID tag readers cover all entrances and core stacking areas, enabling the identification and binding of incoming wires by serial number, specification, and batch.

[0043] After on-site commissioning, the image recognition module achieved an accuracy rate of 96.3%, maintaining an effective recognition rate of approximately 92% even under strong occlusion or partial blur. The system constructed a spatial structure diagram of the wires and accurately mapped the image coordinates to the warehouse's three-dimensional layout model, maintaining a spatial positioning error within ±12cm. The system also achieved a fusion matching accuracy of over 98% for RFID and image data, significantly improving the reliability of wire identification.

[0044] In terms of dynamic inventory management, the system automatically updates approximately 1,200 inventory status records daily. Using anomaly recognition logic, it identified 24 data entries with positional offsets or label discrepancies, automatically marking and isolating them, preventing erroneous calls during subsequent outbound delivery. The path planning module utilizes an obstacle-avoidance path search algorithm. Simulating varying stocking densities, the system achieved an average transport path length optimization rate of 18.7%, while also reducing the transport failure rate from 4.5% with traditional solutions to 1.2%.

[0045] During the outbound delivery process, the system intelligently matched appropriate wires to the requested task and generated handling instructions and routes for the automated transport cart. These instructions included the target wire number, current location, delivery port coordinates, and execution order. During the delivery process, the system dynamically monitored the delivery path through image feedback. The system identified and corrected 37 path deviations without causing any task interruptions, demonstrating the robustness and fault tolerance of its scheduling mechanism.

[0046] Combining the above deployment scenarios and operational results, we further calculated the typical operational indicators of the system during one week of actual operation. The data is detailed in the table below: project Traditional method (control group) System of the present invention (test group) Effect improvement indicators Average daily wire in and out efficiency (roll / day) 137 218 Increased by 59.1% Wire identification accuracy 82.5% 96.3% Increased by 13.8 percentage points Position mapping accuracy (error) ±42cm ±12cm Accuracy improved by 71.4% Label matching accuracy 85.2% 98.1% Increased by 12.9 percentage points Abnormal data identification capability (items / week) Not available 24 articles Significantly enhanced Average path transport optimization rate not applicable 18.7% Significant optimization Failure rate of transport tasks 4.5% 1.2% Reduced by 3.3 percentage points As can be seen from the above-mentioned "Statistical Table of Operational Results of the Intelligent Wire Storage System", the present invention has achieved significant optimization in multiple core indicators compared to traditional management methods. First, in terms of warehousing efficiency, the system has achieved an average daily processing capacity of 218 rolls, an increase of 59.1% compared to the traditional 137 rolls, effectively alleviating the manual bottleneck caused by high-frequency operations and significantly improving the overall operation flow speed. In terms of wire recognition accuracy, this system reached 96.3%, an increase of 13.8 percentage points compared to the 82.5% of manual recognition, and is particularly adapted to the high-precision recognition needs in complex stacking and multi-view scenarios.

[0047] In terms of spatial mapping accuracy, the system controls image coordinate positioning error within ±12cm, while traditional estimation methods have an error of ±42cm, achieving a 71.4% improvement in accuracy. This provides a reliable spatial basis for subsequent path planning and precise handling. In terms of RF and image data fusion, the matching accuracy rate increased from 85.2% to 98.1%, effectively avoiding cable identity mismatches and ensuring the accuracy and consistency of data binding.

[0048] The system also has the ability to automatically identify abnormal data, identifying and isolating 24 abnormal inventory records within a week, a feature that traditional methods lack. This feature improves the system's fault tolerance and data reliability. In terms of path optimization, the intelligent path planning algorithm reduced the average path length by 18.7%, significantly improving equipment scheduling efficiency. Finally, in terms of task execution stability, the system reduced the handling failure rate from 4.5% to 1.2%, ensuring the continuity and reliability of the unmanned operation process.

[0049] In summary, the system has been comprehensively optimized in key links such as intelligent identification, data fusion, dynamic database construction, path scheduling and exception handling, fully demonstrating the practicality, advancement and promotional value of the present invention in complex wire storage environments.

[0050] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for intelligent identification of wire warehouse storage, characterized in that: The steps include: S1. Collect multi-angle image data in the warehouse, use convolutional neural networks to detect wire targets, extract categories, sizes and stacking locations, and generate recognition results; S2. Integrate the recognition results with the warehouse structure diagram, perform spatial mapping, and generate warehouse status data including wire category, location, and hierarchy. S3. Use radio frequency equipment to read the wire electronic tag, obtain batch and specification information, and associate and confirm it with the warehouse status data; S4. Build a dynamic inventory database to record the type, specification, location and status of wires, and continuously update it during the operation process; S5. Receive outbound task requests, analyze target requirements, plan the optimal route based on the inventory database and warehouse status data, and generate transportation instructions; S6. Control the automatic handling equipment to perform the handling operation, and use the image recognition module to monitor in real time. If a deviation is identified, the path replanning is triggered; S7. After the handling is completed, the inventory database is updated, the outbound status and operation log are recorded, and closed-loop management of the entire process is achieved.

2. The intelligent identification method for wire warehouse storage according to claim 1 is characterized in that: The S1 includes: collecting image data through several angle cameras installed in the warehouse, analyzing the images using several pre-trained scale convolutional neural network models, identifying the category, appearance size, position coordinates in the image and stacking level information of each wire, and forming a feature data set of the wire target.

3. The intelligent identification method for wire warehouse storage according to claim 2 is characterized in that: The S2 specifically includes: S21. Based on the obtained target features of each wire rod, extract the corresponding image position coordinates and stacking level information as input data required for spatial positioning; S22. Calling a two-dimensional spatial layout diagram corresponding to the actual structure of the wire warehouse, preliminarily projecting the wire position coordinates in the image to corresponding physical area position points in the warehouse layout diagram; S23. Perform coordinate conversion on the preliminary projection results based on the camera installation parameters, shooting angle, and viewing angle conversion model, completing the transformation operation from the image coordinates to the actual physical coordinates of the warehouse, and generating positioning coordinates consistent with the physical layout; S24. Combining the stacking level information of each wire target with the identified size characteristics, estimating the actual height in the warehouse space, and determining the vertical placement order and overall occupied space range of the wires in the current area; S25. For all wire targets that have completed coordinate transformation and height estimation, spatial position integration is performed to construct a three-dimensional spatial mapping data set, and spatial overlap conflicts and boundary rationality are checked and processed; S26. The processed three-dimensional spatial mapping data is converted into a structured storage status diagram, which includes the category, size, physical location, and stacking level information of each wire.

4. The intelligent identification method for wire warehouse storage according to claim 3 is characterized in that: The S3 specifically includes: S31. Setting up a radio frequency identification reader in the storage environment to scan the wire material entity at the target position for image recognition and extracting the unique identification code, batch number, and specification model from the corresponding electronic tag; S32. Match the identification data extracted from each electronic tag with the wire targets located in the warehouse status structure diagram, perform fusion comparison based on the position coordinates, size information, and category attributes, and generate a matching relationship data table; S33. Establish a one-to-one binding relationship for the items in the generated matching relationship data table whose confidence level is higher than the set threshold, complete the identity confirmation of the image recognition target and the label information, and write the confirmation result into the corresponding wire material entry in the warehouse status structure diagram; S34. For the wire target whose identity has been confirmed, the number field and the specification field in the structure diagram entry are updated to construct a fusion feature set with spatial location, category attributes and identity identification; S35. For wire targets that are not matched successfully or have conflicts, they are uniformly marked as pending review status and output as abnormal data records.

5. The intelligent identification method for wire warehouse storage according to claim 4 is characterized in that: The fusion matching operation of step S32 includes: calculating the matching confidence between the image recognition target and the RFID tag data based on the spatial position coordinates, size characteristics and category attributes of each image recognition target; the confidence is weighted and scored based on the three dimensions of spatial proximity, size similarity and category consistency, and the matching result with the highest confidence is bound as the only corresponding relationship.

6. The intelligent identification method for wire warehouse storage according to claim 5, characterized in that: The S4 specifically includes: S41, extracting the category, specification, unique identification code, physical location, and stacking level information of each wire target based on the fused feature set, and constructing a wire inventory data entry; S42. All inventory data items are uniformly entered into a dynamic inventory database, and a standardized data structure is established for task invocation. Each item record includes the wire number, specification model, category, location coordinates, and stacking level; S43. During the execution of the warehousing operation, status feedback information from the image recognition module is received in real time, and the location coordinates and stacking level fields of the relevant entries in the inventory database are dynamically updated; S44. Perform integrity and consistency checks on the updated inventory data. If any entries with duplicate locations, missing status, or abnormal stacking relationships are found, they are recorded as abnormal data and output to the scheduling system for troubleshooting reference. S45. Output the dynamically updated inventory database to the inbound and outbound scheduling module for use in transportation path planning, task allocation, and resource scheduling steps, thereby forming a real-time and available wire inventory information basis.

7. The intelligent identification method for wire warehouse storage according to claim 6, characterized in that: The S5 specifically includes: S51, receiving a delivery task request, extracting the wire material category, specification, and target quantity information contained in the task as delivery demand input; S52. Retrieve wire records that meet the requirements of the outbound task based on the in-stock items in the inventory database, sort them according to their location coordinates and stacking level priorities, and generate a list of candidate outbound wires. S53. Build a path model based on the warehouse structure diagram and the physical location of the candidate wires, perform path planning operations, and determine the shortest path from the current location of the wires to the delivery point, excluding obstacle areas and high-density stacking areas. S54. After completing the path planning, the feasibility of each path is evaluated and sorted in combination with the execution parameters of the handling equipment to determine the order of outbound delivery and the corresponding handling target location; S55: Generate a handling instruction set, which includes a warehouse-out sequence, a unique number of the corresponding wire material, spatial coordinates, and target position coordinates, and output the instruction set to the automatic handling module.

8. The intelligent identification method for wire warehouse storage according to claim 7 is characterized in that: The path planning in step S53 includes: constructing a dynamic obstacle avoidance graph structure based on the obstacle areas and stacking density information marked in the warehouse structure diagram; performing a path search operation on the graph structure, giving priority to path nodes with lower traffic density, excluding all obstacle areas marked as inaccessible, and obtaining the shortest effective path between the current position of the wire and the delivery point.

9. The intelligent identification method for wire warehouse storage according to claim 8, characterized in that: The S6 specifically includes: controlling the automatic handling equipment to execute the outbound operation one by one according to the handling instruction set, and collecting the current position of the wire and the deviation of the handling path in real time through the image recognition module during the handling process. If a position deviation or path blockage is detected, the path replanning mechanism is immediately triggered to regenerate a feasible path and update the handling instructions.

10. A wire warehouse storage intelligent identification system, applied to a wire warehouse storage intelligent identification method according to any one of claims 1 to 7, characterized in that: Includes the following modules: Image acquisition module, used to obtain multi-angle images of the warehouse and provide image input source; Target recognition module, used to identify wire targets in the image and extract their feature data; Spatial mapping module, used to convert the wire position in the image into physical coordinates of the warehouse; Tag reading module, used to obtain the RFID information of the wire and extract the number and specification; Information fusion module, used to match and bind radio frequency data with image targets; Inventory management module, used to build and update a dynamic inventory database and store structured information; Path planning module, used to generate the optimal transportation path based on inventory and structure diagrams; The instruction generation module is used to output the transport execution instruction including the number and coordinates; The handling control module is used to drive the automatic equipment to perform handling and monitor the task status; The anomaly detection module is used to identify path deviations and data conflicts and trigger replanning.

Images