Tool positioning management and control system and early warning method based on RFID and video fusion

Abstract

The present application relates to the technical field of tool positioning, and discloses a tool positioning management and control system and early warning method based on RFID and video fusion, which comprises the following steps: establishing an association mapping of RFID tag unique code and tool unique identification and constructing a tool information file; configuring an electronic fence based on a management and control area space model, collecting RFID tag signals to perform area attribution inference, generating an RFID side position state and performing first stage state determination; obtaining a monitoring video stream that is spatially registered with the electronic fence space range, performing tool target detection, tool category confirmation and cross-frame target tracking on the video picture to obtain a video side tool existence state; and fusing the RFID side position state and the video side tool existence state to perform second stage determination, and starting a hierarchical early warning strategy to output alarm information according to the determination. The present application reduces the false judgment and missed judgment caused by single sensing.

Description

Technical Field

[0001] This invention relates to the field of tool positioning technology, and more specifically, to a tool positioning control system and early warning method based on RFID and video fusion. Background Technology

[0002] In the fields of intelligent manufacturing, precision machining, and tool management, real-time and reliable location and status monitoring of high-value or high-risk cutting tools is crucial. Currently, the mainstream technologies rely primarily on Radio Frequency Identification (RFID) or video surveillance. RFID-based solutions use a reader network to locate tools by attaching tags, but their signals are easily attenuated, fluctuate, or even lost in complex industrial environments with dense metal, obstructions, or electromagnetic interference, leading to frequent false alarms. For example, temporary signal interference might be mistaken for tool loss or missed detection, or tools that have been illegally removed but whose tags have been damaged might not be detected. Video surveillance solutions, on the other hand, are limited by lighting conditions, shooting angle, background complexity, and the accuracy of the recognition algorithm. When the tool is partially obscured, moving rapidly, or its color is similar to the background, there is a risk of recognition failure or misidentification. Both RFID and video-based solutions share a common drawback: a single sensing dimension. In complex working conditions, they cannot reliably determine and confirm the true status and anomaly types of the tools, resulting in insufficient reliability and practicality of the entire control system. Summary of the Invention

[0003] In view of this, the present invention proposes a tool positioning and control system and early warning method based on RFID and video fusion, aiming to solve the problem that the existing technology cannot reliably identify and confirm the true status and abnormal type of the tool due to reliance on a single sensing method.

[0004] In one aspect, this invention proposes a tool positioning and early warning method based on RFID and video fusion, comprising: Based on the RFID tags on the cutting tools, an association mapping is established between the unique code of the RFID tag and the unique identifier of the cutting tool, and a cutting tool information file is constructed based on the association mapping result. The cutting tool information file includes the cutting tool model, usage rights and life management threshold. Configure electronic fence parameters based on the spatial model of the tool control area to determine the boundary range of the permitted activity area, the warning buffer zone and the prohibited area; collect RFID tag signals based on the RFID reader network, perform area attribution estimation based on the RFID tag signals, and determine the area category of the tool based on the area attribution estimation results to generate the RFID side location status. The first stage of the operation is determined based on the RFID side location status. When the RFID side location status is in the allowed activity area, it is determined to be the first operating state; when the RFID side location status is in the warning buffer area, it is determined to be the second operating state; when the RFID side location status is in the prohibited area or the RFID tag signal is lost, it is determined to be the third operating state. This is used to synchronously acquire and spatially register the monitoring video stream with the electronic fence spatial range, process the monitoring video stream, perform tool target detection on the video screen, and confirm the tool category and track cross-frame targets of the candidate targets obtained by tool target detection to obtain the tool presence status on the video side. A second-stage fusion determination is performed based on the RFID-side location status and the video-side tool presence status to confirm the tool anomaly type. Based on the type of anomaly, the corresponding hierarchical early warning strategy is activated, and the early warning information is output as at least one of local audible and visual alarms, management platform alarms, or mobile terminal alarms.

[0005] Furthermore, when establishing the association mapping between the unique code of the RFID tag and the unique identifier of the tool based on the RFID tag on the tool, it includes: Obtain the knives to be included in the management and assign a unique knife identifier to each knife; Read the unique encoding information of the RFID tag attached to the cutting tool, and perform format consistency and repeatability checks on the unique encoding information; Each verified RFID tag's unique code is bound to its corresponding unique tool identifier, forming a unique correspondence. The unique correspondence is stored and marked to obtain the association mapping result between the unique code of the RFID tag and the unique identifier of the tool.

[0006] Furthermore, when constructing the tool information archive based on the association mapping results, it includes: Based on the association mapping result, a corresponding tool file index identifier is generated, and the tool file index identifier is associated with the tool unique identifier; The tool model information, usage permission information, and life management threshold information are obtained based on the tool's unique identifier, and the consistency of the tool model information, usage permission information, and life management threshold information is verified. When the consistency check passes, the model information, usage permission information and life management threshold information are written into the file record corresponding to the tool file index identifier; If the consistency check fails, the corresponding tool file index is marked as abnormal. The tool information file is obtained by integrating the model information, usage permission information and life management threshold information that have passed the consistency verification.

[0007] Furthermore, when configuring the electronic fence parameters based on the spatial model of the tool control area, the following are included: Acquire spatial structure information of the tool control area, including the area boundary contour, entrance and exit locations, and passable path information; The boundary contour of the tool control area is analyzed, and a closed boundary line is formed according to the boundary contour to obtain the basic fence boundary. Based on the basic fence boundary, the fence extends outward in sequence according to the preset spatial offset rules to form a multi-layered boundary structure of permitted activity area, early warning buffer area and prohibited area. The multi-layer boundary structure is converted into a computable set of electronic fence parameters; The region discrimination rules are determined based on the electronic fence parameter set; The integrity and consistency of the electronic fence parameter set are verified, and the electronic fence parameters are configured when the verification passes, thus obtaining the electronic fence parameters.

[0008] Furthermore, when collecting RFID tag signals and generating RFID-side location status based on the RFID reader network, the process includes: Within the tool control area, multiple RFID readers deployed at different spatial locations collect signal data corresponding to RFID tags within the same sampling period. The signal data includes signal strength information and signal continuity information received by each RFID reader. Based on the spatial location and signal strength information of each RFID reader, the RFID tag signal is assigned a region, and the estimated region assignment result of the RFID side is obtained. A consistency analysis is performed on the RFID-side area attribution estimation results obtained within a continuous sampling period. If the area determination results corresponding to the RFID-side area attribution estimation results remain consistent within a continuous sampling period, the RFID-side area attribution estimation results are determined to be valid area estimation results. The effective area estimation results are compared with the permitted activity area, the warning buffer area and the prohibited area to determine the area category where the tool is located; If no valid area estimation result satisfying the consistency condition is obtained within a continuous sampling period, the RFID tag is determined to be in a signal loss state. Generate RFID-side location status based on region category or signal loss status.

[0009] Furthermore, when performing the first-stage state determination based on the RFID-side location status, it includes: When the region determination result switches across regions within a continuous sampling period, the tool is determined to be in a boundary determination state. Under the boundary determination state, the corresponding region determination results within the continuous sampling period are summarized and analyzed to determine the region category with the most consecutive occurrences as the boundary determination result. When the boundary determination result corresponds to the allowed activity area, the tool is confirmed to be in the first operating state; When the boundary determination result corresponds to the early warning buffer area, the tool is confirmed to be in the second operating state; When the boundary determination result corresponds to a prohibited area or a signal loss state occurs in the area determination result, the tool is confirmed to be in the third operating state. The confirmed operating status is output as the first-stage status determination result.

[0010] Furthermore, when acquiring the presence status of the tool on the video side, the following includes: Acquire surveillance video streams that are spatially registered with the electronic fence's spatial range, and process the surveillance video streams within a continuous video sampling period; Within each video sampling period, tool target detection is performed on the video frame, and the candidate targets obtained from the tool target detection are confirmed as tool types and tracked across frames to obtain the tool detection results; A consistency analysis is performed on the tool detection results obtained within a continuous video sampling period. If the tool detection results remain consistent within the continuous video sampling period, the tool existence state is determined to be a valid existence state. If no target matching the appearance characteristics of the tool is detected within a continuous video sampling period, or if the tool detection result changes frequently within a continuous video sampling period, the tool's existence state is determined to be an uncertain state. Based on the valid or uncertain state, the video-side tool presence state is generated.

[0011] Furthermore, the second-stage fusion determination based on the RFID-side location status and the video-side tool presence status includes: Obtain the RFID-side location status and the video-side tool presence status within the same time-related window, and construct the corresponding status combination; When the RFID-side location status and the video-side tool presence status remain consistent within a continuous time correlation window, the tool is confirmed to be in the corresponding fusion confirmation state. When the RFID-side position status indicates that the tool is in the second or third operating state, and the video-side tool presence status is valid, it is confirmed that the tool is in a high-risk fusion state. When the RFID-side position status indicates that the tool is in the first operating state, and the video-side tool existence status is uncertain, the tool is confirmed to be in the pending confirmation fusion state. When the RFID side position status is signal loss and the video side tool presence status is valid, it is confirmed that the tool is in an abnormal fusion state. The fusion confirmation status, high-risk fusion status, pending confirmation fusion status, or abnormal fusion status are output as the second-stage fusion judgment results.

[0012] Furthermore, when initiating a tiered early warning strategy based on the second-stage fusion determination result, it includes: When the fusion determination result is a fusion confirmation state, the current monitoring state is maintained and no warning is triggered; When the fusion determination result is a fusion status to be confirmed, a low-level early warning strategy is activated to continuously track the corresponding tool and increase the sampling frequency of the RFID-side position status and the video-side tool presence status. When the fusion determination result is a high-risk fusion state, an intermediate early warning strategy is activated to generate abnormal early warning information for the corresponding tool and push the abnormal early warning information to the management and control platform. When the fusion determination result is an abnormal fusion state, an advanced early warning strategy is activated, an emergency early warning message is generated, and a multi-channel alarm mechanism is triggered simultaneously. Based on the execution results of the tiered early warning strategy, the corresponding early warning event information is recorded to form an early warning processing record.

[0013] Compared with existing technologies, the advantages of this invention are as follows: By associating RFID tags with unique tool identifiers and constructing tool information files containing tool model, usage rights, and lifespan management thresholds, the identity and management attributes of tools during the control process can be uniformly identified and invoked, thereby avoiding management blind spots caused by extensive control based solely on physical location; by configuring electronic fence parameters based on the tool control area spatial model and combining RFID reader network to estimate the area affiliation of RFID tag signals, real-time determination of the area category where the tool is located can be achieved, enabling the tool's position status to be finely distinguished into permitted activity, warning buffer, and prohibited areas, improving the effectiveness and controllability of spatial control; by setting a first-stage status determination mechanism, the RFID-side position status is further mapped to different operating states. The system establishes a clear operational semantics for the area determination results, providing a clear foundation for subsequent risk identification. By introducing monitoring video streams spatially registered with the electronic fence's spatial range, the system detects knife targets in the video footage. Combined with knife category confirmation and cross-frame target tracking, it establishes the knife presence status on the video side, thus providing effective status judgment even under conditions of occlusion, lighting changes, or RFID signal anomalies. A second-stage fusion judgment is performed on the RFID-side location status and the video-side knife presence status to achieve cross-verification of multi-source information, reducing the risk of misjudgment or omission caused by a single sensing method. Based on the fusion judgment results, a tiered early warning strategy is initiated, matching the early warning output with the knife anomaly type. This ensures control safety while avoiding unnecessary frequent alarms, improving the reliability and practicality of knife location control and anomaly early warning.

[0014] On the other hand, this application also provides a tool positioning and control system based on RFID and video fusion, used to implement the above-mentioned tool positioning and early warning method based on RFID and video fusion, including: The tool information association module is used to establish an association mapping between the unique code of the RFID tag and the unique identifier of the tool based on the RFID tag on the tool, and to construct a tool information file based on the association mapping result. The tool information file includes the tool model, usage rights and life management threshold. Configure electronic fence parameters based on the spatial model of the tool control area to determine the boundary range of the permitted activity area, the warning buffer zone and the prohibited area; collect RFID tag signals based on the RFID reader network, perform area attribution estimation based on the RFID tag signals, and determine the area category of the tool based on the area attribution estimation results to generate the RFID side location status. The RFID operation status determination module is used to determine the first stage of the status based on the RFID side location status. When the RFID side location status is in the allowed activity area, it is determined to be in the first operation status; when the RFID side location status is in the warning buffer area, it is determined to be in the second operation status; when the RFID side location status is in the prohibited area or when the RFID tag signal is lost, it is determined to be in the third operation status. The video-side knife recognition module is used to synchronously acquire monitoring video streams that are spatially registered with the electronic fence spatial range, process the monitoring video streams, perform knife target detection on the video images, and confirm the knife category and track cross-frame targets of the candidate targets obtained by the knife target detection to obtain the knife presence status on the video side. The fusion status determination module is used to perform a second-stage fusion determination based on the RFID-side location status and the video-side tool presence status to confirm the tool abnormality type. The strategy execution module is used to activate the corresponding hierarchical early warning strategy according to the anomaly type, and output the early warning information as at least one of local audible and visual alarms, management platform alarms, or mobile terminal alarms.

[0015] It is understandable that the aforementioned tool positioning and control system and early warning method based on RFID and video fusion have the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0016] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart illustrating a tool positioning and early warning method based on RFID and video fusion provided in an embodiment of the present invention; Figure 2 This is a functional block diagram of a tool positioning and control system based on RFID and video fusion provided in an embodiment of the present invention. Detailed Implementation

[0017] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0018] See Figure 1 As shown, this application proposes a tool positioning and early warning method based on RFID and video fusion, including: S1: Based on the RFID tags on the cutting tools, establish an association mapping between the unique code of the RFID tag and the unique identifier of the cutting tool, and construct a cutting tool information file based on the association mapping result. The cutting tool information file includes the cutting tool model, usage rights and life management threshold. S2: Configure electronic fence parameters according to the spatial model of the tool control area, and determine the boundary range of the allowed activity area, the warning buffer zone and the prohibited area; collect RFID tag signals based on the RFID reader network, perform area attribution estimation based on the RFID tag signals, and determine the area category of the tool based on the area attribution estimation results, and generate the RFID side location status. S3: The first stage of the status determination is based on the RFID side location status. When the RFID side location status is in the allowed activity area, it is determined to be the first operating state; when the RFID side location status is in the warning buffer area, it is determined to be the second operating state; when the RFID side location status is in the prohibited area or the RFID tag signal is lost, it is determined to be the third operating state. S4: Used to synchronously acquire the monitoring video stream that has been spatially registered with the electronic fence spatial range, process the monitoring video stream, perform tool target detection on the video screen, and confirm the tool type and track cross-frame targets of the candidate targets obtained by tool target detection to obtain the tool presence status on the video side. S5: Perform a second-stage fusion judgment based on the RFID side location status and the video side tool presence status to confirm the tool anomaly type; S6: Activate the corresponding hierarchical early warning strategy according to the anomaly type, and output the early warning information as at least one of local audible and visual alarms, management platform alarms, or mobile terminal alarms.

[0019] Specifically, passive RFID tags are attached to the cutting tools. These tags carry identification information to distinguish different tools. The unique RFID code within the tag serves as the basis for tool identification during method execution. By establishing a one-to-one mapping between the unique RFID tag code and a pre-assigned unique tool identifier, each tool has a unique and traceable identity during method execution. The assignment of the unique tool identifier is based on the tool's actual management number, warehousing record, or manufacturing number, ensuring that the unique tool identifier is not duplicated throughout the entire control scope. After completing the mapping between the RFID tag's unique code and the tool's unique identifier, a tool information file is constructed based on the mapping results. This file centrally records basic attribute information related to the tool. Tool model information is determined based on the tool's specifications or design purpose; usage permissions are set based on the applicable work procedures, personnel categories, or management systems; and lifespan management thresholds are determined based on historical usage data, the manufacturer's recommended usage period, or statistical usage data of similar tools, characterizing the tool's safe usage limits or effective usage period. After establishing the tool identification and attribute information, a corresponding spatial model is created based on the tool control area where the tool may be active or stored. This spatial model consists of the boundary contour, entrance / exit locations, and accessible paths. The spatial structure information is derived from on-site surveying results, building plans, or digital site models. Based on this, electronic fence parameters are configured according to the spatial model of the tool control area. The configuration process includes analyzing the boundary contour of the tool control area to form a closed boundary line, and then generating a multi-layered spatial boundary structure based on the closed boundary line according to preset spatial offset rules. The permitted activity area corresponds to the space range where the tool can normally operate or be stored. The warning buffer area is formed by extending the permitted activity area outwards or inwards, used to trigger risk warnings in advance when the tool approaches the risk boundary. The prohibited area limits the space range where the tool cannot enter. The spatial offset rules are determined based on the tool's movement range during actual operation, safety distance requirements, and on-site management regulations, thus ensuring that the electronic fence parameters match the actual usage scenario. During execution, passive RFID tag signals are collected by RFID readers deployed within the tool control area. Combined with the spatial location information of the RFID readers within the tool control area, as well as the signal strength and continuity information of the passive RFID tag signals, the area attribution of the passive RFID tag signals is calculated to obtain the RFID-side area attribution estimation result. Based on the RFID-side area attribution estimation result, the area category of the tool is determined. The area category determination is based on the area attribution relationship between the RFID-side area attribution estimation result and the permitted activity area, warning buffer area, and prohibited area, thereby generating the RFID-side location status.Based on the RFID-side location status, a first-stage state determination is performed to distinguish between three operating states: a first operating state corresponding to the allowable activity area, a second operating state corresponding to the warning buffer area, and a third operating state corresponding to the prohibited area or the loss of passive RFID tag signals. The determination of passive RFID tag signal loss is based on the situation where no valid passive RFID tag signal is received during continuous sampling. To supplement the area attribution determination results obtained based on passive RFID tag signals, a monitoring video stream spatially registered with the electronic fence spatial range is simultaneously acquired. Spatial registration is completed based on the correspondence between the installation position and shooting angle of the monitoring camera and the electronic fence spatial model. During the processing of the monitoring video stream, knife target detection is performed on the video image, and the candidate targets obtained from the knife target detection are confirmed for knife category and tracked across frames, thereby generating the video-side knife presence status. After acquiring the RFID-side location status and the video-side knife presence status, a second-stage fusion determination is performed based on the RFID-side location status and the video-side knife presence status. By comprehensively analyzing the passive RFID-side area attribution determination results and the video-side identification results, it is confirmed whether there are any abnormalities in the knife and the corresponding abnormality type. Finally, based on the confirmed anomaly type, the corresponding graded early warning strategy is activated. The graded early warning strategy is determined according to the risk level of the anomaly type. The early warning information is output in at least one of the following ways: local audible and visual alarm, management platform alarm, or mobile terminal alarm, so as to realize continuous control of tool position and timely early warning of abnormal situations.

[0020] In one specific embodiment, a knife already under control is selected as an example object. A passive RFID tag is fixedly installed on the knife. The unique code of the passive RFID tag has been associated with the unique identifier of the knife, and a corresponding knife information file is established based on the association mapping result. The knife information file records the knife model, usage rights, and lifespan management threshold. Based on the actual spatial layout of the work site where the knife is located, a spatial model of the knife control area is pre-constructed, and electronic fence parameters are configured based on the spatial model. The electronic fence parameters divide the knife control area into a permitted activity area, a warning buffer area, and a prohibited area. During the method execution, when the knife is within the permitted activity area, the passive RFID tag signal is collected by an RFID reader, and the area affiliation is estimated. The generated RFID-side location status corresponds to the permitted activity area. Based on the RFID-side location status, a first-stage status determination is performed, confirming that the knife is in the first operating state. Subsequently, when the knife gradually approaches the control boundary and enters the warning buffer area during actual use or handling, the RFID-side location status changes, and the first-stage status determination result is updated to the second operating state. At this point, the method enters the stage of focusing on the knife's status. Furthermore, when the tool continues to move and enters a prohibited area, or when a passive RFID tag signal is lost during movement, the RFID-side position status is determined to be an abnormal position status, and the first-stage status determination result is updated to the third operating state. Simultaneously, a monitoring video stream spatially registered with the electronic fence is acquired, processed, and tool target detection is performed on the video footage. The candidate targets obtained from the tool target detection are then confirmed for tool type and tracked across frames, forming a video-side tool presence status. In this embodiment, the video-side tool presence status is a valid presence status. Subsequently, a second-stage fusion determination is performed based on the RFID-side position status and the video-side tool presence status to comprehensively determine that the tool has crossed the boundary abnormally and confirm the corresponding abnormality type. Finally, based on the confirmed abnormality type, a corresponding hierarchical early warning strategy is activated, outputting the early warning information as at least one of a local audible and visual alarm, a management platform alarm, or a mobile terminal alarm, thereby completing the location control and abnormal early warning of the tool crossing the boundary.

[0021] In some embodiments of this application, when establishing an association mapping between the unique code of the RFID tag and the unique identifier of the tool based on the RFID tag on the tool, the following steps are included: Obtain the knives to be included in the management and assign a unique knife identifier to each knife; Read the unique coding information of the RFID tag attached to the cutting tool, and perform format consistency and duplicate verification on the unique coding information; Each verified RFID tag's unique code is bound to its corresponding unique tool identifier, forming a unique correspondence. The unique correspondence is stored and marked to obtain the association mapping result between the unique code of the RFID tag and the unique identifier of the tool.

[0022] Specifically, the process involves selecting knives requiring location and early warning control from the actual management scope, and generating a unique identifier for each knife within the control scope. This unique identifier is generated based on existing numbering rules in the inventory management ledger, production equipment number, or manual registration records to ensure no duplication of unique identifiers between different knives. Next, the unique RFID tag code information fixed on the knife is read and compared with pre-defined encoding rules. These rules limit the length range, character type composition, and encoding structure order of the unique RFID tag code. A step-by-step comparison is performed to determine if the unique RFID tag code meets the encoding format requirements, thus completing the format consistency check. After format consistency verification, the read unique RFID tag code is compared with all stored unique RFID tag code records. If no identical code exists in the existing records, the unique RFID tag code is determined not to be reused, thus completing the duplicate check. If a unique RFID tag code is detected to match any code in an existing record, a duplicate is determined, and the current association mapping process is terminated. With the RFID tag's unique code passing both format consistency and duplicate verification, the unique RFID tag code is bound one-to-one with the corresponding unique tool identifier, forming a unique mapping relationship. Subsequently, this one-to-one mapping relationship is written into a dataset used to record tool identification information, and a valid marker is added to the mapping relationship. This valid marker indicates that the correspondence between the RFID tag's unique code and the tool's unique identifier has been established and can be used in subsequent positioning and determination processes, ultimately yielding the association mapping result between the RFID tag's unique code and the tool's unique identifier.

[0023] In some embodiments of this application, constructing a tool information file based on the association mapping result includes: Generate a corresponding tool file index identifier based on the association mapping result, and associate the tool file index identifier with the tool unique identifier; The tool model information, usage permission information and life management threshold information are obtained based on the tool's unique identifier, and the consistency of the model information, usage permission information and life management threshold information is verified. When the consistency check passes, the model information, usage permission information, and life management threshold information are written into the file record corresponding to the tool file index identifier; If the consistency check fails, the corresponding tool file index is marked as abnormal. The tool information file is obtained by integrating the model information, usage permission information and life management threshold information that have passed the consistency verification.

[0024] Specifically, in the process of constructing tool information archives based on the association mapping results, firstly, based on the established association mapping results between the unique codes of RFID tags and the unique identifiers of tools, a corresponding tool archive index identifier is generated for each set of association mapping results. The tool archive index identifier is used to uniquely locate a tool archive record during the method execution process, and an association relationship is established between the tool archive index identifier and the corresponding unique tool identifier, thereby ensuring a one-to-one correspondence between the tool archive index identifier and the specific tool. After the tool archive index identifier is generated, tool model information, usage permission information, and life management threshold information are obtained from the pre-established tool basic information source based on the unique tool identifier. Among them, the tool model information is used to describe the specification type or structural category of the tool, the usage permission information is used to limit the scope of operations or usage conditions in which the tool is allowed to participate, and the life management threshold information is used to characterize the safe or effective use limits of the tool during use. After obtaining the tool model information, usage permission information, and lifespan management threshold information, a consistency verification process is performed on these three information. This process includes comparing the tool model information with the registered model record corresponding to the tool's unique identifier to determine if the tool model information matches the tool type corresponding to the unique identifier; matching the usage permission information with the permission configuration rules in the current method execution environment to determine if the usage permission information meets the operating conditions allowed for tool use; and further comparing the lifespan management threshold information with historical usage records or preset lifespan management rules for the same type of tool to determine if the lifespan management threshold information is within a reasonable range. When the tool model information, usage permission information, and lifespan management threshold information all pass the consistency verification process, they are written into the file record corresponding to the tool file index identifier, thus completing the effective registration of the tool file information. If any of the tool model information, usage permission information, or lifespan management threshold information fails the consistency check, the corresponding tool file index is marked as abnormal. During method execution, subsequent location and determination processes related to this tool file index are blocked to prevent abnormal file information from participating in subsequent status determinations. Finally, the tool model information, usage permission information, and lifespan management threshold information that have passed the consistency check are integrated to form a complete tool information file.

[0025] In some embodiments of this application, configuring electronic fence parameters based on a spatial model of the tool control area includes: Acquire spatial structure information of the tool control area, including the area boundary outline, entrance and exit locations, and passable path information; The boundary contour of the tool control area is analyzed, and a closed boundary line is formed according to the boundary contour to obtain the basic fence boundary. Based on the basic fence boundary, the fence extends outward in sequence according to the preset spatial offset rules, forming a multi-layered boundary structure of permitted activity area, early warning buffer area and prohibited area; Convert the multi-layered boundary structure into a computable set of electronic fence parameters; Determine the area discrimination rules based on the electronic fence parameter set; Perform integrity and consistency checks on the electronic fence parameter set, and complete the configuration of the electronic fence parameters when the checks pass, thus obtaining the electronic fence parameters.

[0026] Specifically, in configuring electronic fence parameters based on the spatial model of the tool control area, the spatial structure information of the tool control area is first acquired. This information describes the spatial range involved in the actual use or storage of the tool. The spatial structure information includes the area boundary outline, the entrance and exit locations within the area, and the permissible paths for the tool within the area. The area boundary outline can be derived from on-site surveying results, building plans, or a digital site model, and is used to define the external boundary of the tool control area. The entrance and exit locations identify key nodes for the tool to enter and exit the area. The permissible path information describes the actual routes the tool may take within the area. After acquiring the complete spatial structure information of the tool control area, the area boundary outline is parsed. By connecting continuous boundary points in the outline, a closed boundary line is formed, thus obtaining the basic fence boundary used to describe the outer boundary of the tool control area.

[0027] After obtaining the basic fence boundary, it is used as a reference to expand outward layer by layer according to preset spatial offset rules, generating a multi-layered boundary structure corresponding to the permitted activity area, the warning buffer area, and the prohibited area. The preset spatial offset rules are used to define the spatial interval relationship between each layer boundary. They are determined based on the movement range of the tool during actual use, the personnel operation safety distance requirements, and on-site management regulations, thereby ensuring a clear spatial distinction between the boundaries of different areas. By executing the preset spatial offset rules, the innermost first boundary layer, representing the permitted activity area, is first formed; then, based on the first boundary layer, it expands outward to form the second boundary layer, representing the warning buffer area; finally, based on the second boundary layer, it expands outward to form the outermost third boundary layer, representing the prohibited area. This completes the construction of a multi-layered boundary structure corresponding to the permitted activity area, the warning buffer area, and the prohibited area, nested sequentially from the inside out.

[0028] After constructing the multi-layer boundary structure, it is converted into a computable electronic fence parameter set. This parameter set transforms spatial boundary information into parameters usable for position determination during method execution. The electronic fence parameter set includes boundary coordinates describing the spatial extent of each boundary layer and the spatial relationships between them. After obtaining the electronic fence parameter set, region discrimination rules are determined based on it. These rules determine the region category of the tool based on the spatial relationship between the tool's current position and each boundary layer during method execution. Subsequently, integrity and consistency checks are performed on the electronic fence parameter set. Integrity checks determine if the parameter set contains all boundary information corresponding to permitted activity areas, warning buffer areas, and prohibited areas. Consistency checks determine if the spatial relationships between boundary layers conform to the constraints of preset spatial offset rules. If the electronic fence parameter set passes both integrity and consistency checks, the electronic fence parameters are configured, ultimately yielding the parameters used for subsequent tool position determination.

[0029] In some embodiments of this application, when collecting RFID tag signals and generating RFID-side location status based on an RFID reader network, the following steps are included: Within the tool control area, multiple RFID readers deployed at different spatial locations collect signal data corresponding to RFID tags within the same sampling period. The signal data includes signal strength information and signal continuity information received by each RFID reader. Based on the spatial location and signal strength information of each RFID reader, the RFID tag signal is assigned a region, and the estimated region assignment result of the RFID side is obtained. A consistency analysis is performed on the RFID-side area attribution estimation results obtained within a continuous sampling period. If the area determination results corresponding to the RFID-side area attribution estimation results remain consistent within a continuous sampling period, the RFID-side area attribution estimation results are determined to be valid area estimation results. The effective area estimation results are compared with the permitted activity area, the warning buffer area and the prohibited area to determine the area category where the tool is located; If no valid area estimation result satisfying the consistency condition is obtained within a continuous sampling period, the RFID tag is determined to be in a signal loss state. Generate RFID-side location status based on region category or signal loss status.

[0030] Specifically, in the process of collecting RFID tag signals and generating RFID-side location status based on an RFID reader network, multiple RFID readers are first deployed within the tool control area. These readers are fixed in different spatial locations, allowing for multi-point coverage of the passive RFID tags on the same tool. Within the same sampling period, multiple RFID readers synchronously read the passive RFID tag signals, acquiring the corresponding signal data. This signal data includes the signal strength information received by each RFID reader and the signal continuity information of the passive RFID tag signal within the sampling period. The signal strength information reflects the relative proximity between the passive RFID tag and the RFID reader, while the signal continuity information reflects whether the passive RFID tag signal can be continuously read within the sampling period. Subsequently, a region attribution calculation is performed based on the spatial location and signal strength information of the RFID readers. This process includes normalizing and sorting the signal strength information of multiple RFID readers to obtain a set of RFID readers that correspond to the relative proximity of the passive RFID tags. The process also includes obtaining the set of coverage units of the passive RFID tags in the current sampling period based on the correspondence between the RFID reader set and the RFID reader coverage units. Furthermore, the process includes obtaining the RFID-side region attribution estimation result based on the spatial correspondence between the set of coverage units and the boundaries of permitted activity areas, warning buffer areas, and prohibited areas, thereby avoiding the conversion of passive RFID tag signals into precise coordinate information. After obtaining the RFID-side area attribution estimation results, a consistency analysis is performed on the RFID-side area attribution estimation results obtained within continuous sampling periods. This consistency analysis includes mapping the RFID-side area attribution estimation results corresponding to each sampling period within the continuous sampling period to area determination results, and determining whether the area determination results within the continuous sampling periods are consistent. When the area determination results are consistent within the continuous sampling periods, it is confirmed that the RFID-side area attribution estimation results have a stable area attribution relationship, and the RFID-side area attribution estimation results are determined as valid area estimation results. Subsequently, the valid area estimation results are compared with the permitted activity area, the warning buffer area, and the prohibited area to determine the area inclusion relationship. This determination includes determining whether the set of coverage units corresponding to the valid area estimation results falls within the boundary range of the permitted activity area, the boundary range of the warning buffer area, or the boundary range of the prohibited area, thereby determining the area category of the tool. If no valid area estimation result satisfying the consistency condition is obtained within the continuous sampling period, the signal continuity information collected by the RFID reader within the continuous sampling period is used to determine whether the passive RFID tag signal is continuously unreadable. If it is continuously unreadable, the passive RFID tag is determined to be in a signal loss state.Finally, the RFID-side location status is generated based on the region category or signal loss status. The RFID-side location status is used to characterize the region affiliation status or signal anomaly status of the tool during the current sampling period.

[0031] In some embodiments of this application, the first-stage state determination based on the RFID-side location status includes: When the region determination result switches across regions within a continuous sampling period, the tool is determined to be in the boundary determination state. In the boundary determination state, the corresponding region determination results within the continuous sampling period are summarized and analyzed to determine the region category with the most consecutive occurrences as the boundary determination result. When the boundary determination result corresponds to the allowed active area, the tool is confirmed to be in the first operating state; When the boundary determination result corresponds to the warning buffer area, the tool is confirmed to be in the second running state; When the boundary determination result corresponds to a prohibited area or a signal loss state occurs in the area determination result, the tool is confirmed to be in the third running state. The confirmed operating status is output as the first-stage status determination result.

[0032] Specifically, based on the obtained valid position results, the region determination results corresponding to each sampling period within a continuous sampling period are processed using a time series analysis. The region determination results indicate which type of region the tool falls into within the allowed activity region, warning buffer region, or prohibited region within the corresponding sampling period. When the region determination results of adjacent sampling periods switch between the allowed activity region, warning buffer region, and prohibited region within a continuous sampling period, it indicates that the tool's current position is near the boundary of different regions, and the spatial position calculation results are affected by the boundary proximity effect or signal fluctuations. At this time, the tool is determined to have entered the boundary determination state. After entering the boundary determination state, the region determination results of a single sampling period are not directly used. Instead, multiple region determination results obtained within a continuous sampling period are summarized and analyzed. The summary analysis includes counting the occurrence frequency of each region category within a continuous sampling period, and determining the region category with the most occurrences as the boundary determination result, thereby reducing the impact of instantaneous position jitter or signal fluctuations on the operation status determination. After obtaining the boundary determination results, these results are mapped to the operating status rules. When the boundary determination result corresponds to a permitted activity area, the tool is confirmed to be in the first operating state, indicating that the tool is within the permitted use or normal storage range. When the boundary determination result corresponds to a warning buffer area, the tool is confirmed to be in the second operating state, indicating that the tool is approaching a restricted area and needs to enter the warning monitoring state. When the boundary determination result corresponds to a prohibited area or an RFID tag signal loss occurs in the area determination results within a continuous sampling period, the tool is confirmed to be in the third operating state, indicating that the tool is in a high-risk position or there is a positioning anomaly. Finally, the confirmed first, second, or third operating state is output as the first-stage status determination result.

[0033] In some embodiments of this application, obtaining the presence status of the tool on the video side includes: Acquire surveillance video streams that are spatially registered with the electronic fence's spatial range, and process the surveillance video streams within a continuous video sampling period; Within each video sampling period, tool target detection is performed on the video frame, and the candidate targets obtained from the tool target detection are confirmed as tool types and tracked across frames to obtain the tool detection results; A consistency analysis is performed on the tool detection results obtained within a continuous video sampling period. If the tool detection results remain consistent within the continuous video sampling period, the tool existence state is determined to be a valid existence state. If no target matching the appearance characteristics of the tool is detected within a continuous video sampling period, or if the tool detection result changes frequently within a continuous video sampling period, the tool's existence state is determined to be an uncertain state. Based on the valid or uncertain state, the video-side tool presence state is generated.

[0034] Specifically, spatial registration is performed between the surveillance video stream and the electronic fence spatial range. This registration is based on a spatial model of the tool control area, which describes the boundaries of permitted activity areas, warning buffer areas, and prohibited areas. Simultaneously, installation position parameters, shooting direction parameters, and imaging angle parameters corresponding to the surveillance cameras are established. By mapping pixel areas in the surveillance image to the boundary ranges of the tool control area, a regional correspondence is formed between the video image and the permitted activity areas, warning buffer areas, and prohibited areas. This ensures that identified tool targets in the video image can be associated with the region category of the permitted activity area, warning buffer area, or prohibited area. After spatial registration, the surveillance video stream is processed according to the video sampling period. The video sampling period is determined based on the frame rate of the surveillance cameras and the control response requirements. At least one frame of video is extracted within each video sampling period as the analysis object. Within each video sampling period, tool target detection is performed on the captured video frame to obtain a set of candidate tool targets. Tool category confirmation is then performed on this set to determine whether the candidate targets satisfy the tool category feature constraints and output the confirmation result. Cross-frame target tracking is then performed on the candidate targets that pass the tool category confirmation. This tracking establishes the association between the same tool target within consecutive video sampling periods and obtains a consistent target trajectory, thus yielding the tool detection result for the corresponding video sampling period. Tool category feature constraints are determined based on a pre-established set of tool appearance features, which includes at least one of the tool's contour features, size ratio features, edge features, and surface reflection characteristics. This set is used to reduce the risk of false detections and missed detections caused by occlusion, lighting changes, or differences in tool shape. The tool detection result characterizes whether a tool target was detected within the corresponding video sampling period and the tool target's location within the video frame. Subsequently, a consistency analysis is performed on the tool detection results obtained within the continuous video sampling period. This analysis includes comparing whether the tool detection results persist and whether the tool target trajectory consistency results remain stable within the continuous video sampling period. When the tool detection results remain consistent within the continuous video sampling period, the tool presence state is determined to be a valid presence state, indicating that the video side can continuously confirm the existence of the tool target. When no target matching the tool's appearance characteristics is detected within the continuous video sampling period, or the tool detection results change frequently within the continuous video sampling period, or the tool target trajectory consistency results cannot form a stable correlation, the tool presence state is determined to be an uncertain state, indicating that the video side cannot stably confirm the existence of the tool target. Finally, the video-side tool presence state is generated based on the valid or uncertain presence state.

[0035] In some embodiments of this application, the second-stage fusion determination based on the RFID-side location status and the video-side tool presence status includes: Obtain the RFID-side location status and the video-side tool presence status within the same time-related window, and construct the corresponding status combination; When the RFID-side location status and the video-side tool presence status remain consistent within a continuous time correlation window, the tool is confirmed to be in the corresponding fusion confirmation status. When the RFID-side position status indicates that the tool is in the second or third operating state, and the video-side tool presence status is valid, the tool is confirmed to be in a high-risk fusion state. When the RFID side position status indicates that the tool is in the first operating state, and the video side tool existence status is uncertain, the tool is confirmed to be in the pending confirmation fusion state. When the RFID side location status is signal loss and the video side tool presence status is valid, it is confirmed that the tool is in an abnormal fusion state. The fusion confirmation status, high-risk fusion status, pending confirmation fusion status, or abnormal fusion status will be output as the second-stage fusion judgment result.

[0036] Specifically, in the second-stage fusion determination process based on the RFID-side location status and the video-side tool presence status, a time alignment mechanism is first established to obtain the RFID-side location status and video-side tool presence status within the same time association window. The time association window is used to limit the time range for the RFID-side location status and video-side tool presence status to participate in the fusion determination. The time association window can be determined based on the timestamp alignment relationship between the RFID sampling period and the video sampling period, thereby ensuring that the RFID-side location status and video-side tool presence status correspond to the tool status within the same time period. After time alignment is completed, the RFID-side location status and video-side tool presence status within the time association window are extracted and a state combination is constructed. The state combination describes the joint value relationship between the RFID-side location status and video-side tool presence status within the same time association window. After obtaining the state combination, a consistency verification process for consecutive time association windows is performed. The consistency verification process includes comparing whether the state combinations within adjacent time association windows maintain the same combination type. When the state combinations remain consistent within consecutive time association windows, the tool is confirmed to be in a fusion confirmation state. The fusion confirmation state indicates that the RFID-side location status and video-side tool presence status form a stable and consistent joint determination relationship. Beyond the fusion confirmation status determination, further risk combination classification and identification are performed. When the RFID-side location status indicates the tool is in the second or third operating state, and the video-side tool presence status is valid, the tool is confirmed to be in a high-risk fusion state. A high-risk fusion state indicates that the tool has approached a restricted boundary or entered a restricted area, and the video-side identification result confirms the tool target's existence, thus meeting the high-risk handling conditions. When the RFID-side location status indicates the tool is in the first operating state, and the video-side tool presence status is uncertain, the tool is confirmed to be in a pending confirmation fusion state. This indicates that the RFID-side location status indicates the tool is still within the permitted activity area, but the video-side identification result cannot reliably confirm the tool target's existence, requiring a review or continuous tracking process. When the RFID-side location status is in a signal loss state, and the video-side tool presence status is valid, the tool is confirmed to be in an abnormal fusion state. This abnormal fusion state indicates that there is a signal anomaly on the RFID side, but the video side confirms the tool target's existence, belonging to an abnormal combination requiring key handling. In addition to the above state combinations, the second-stage fusion determination also includes processing rules for other state combinations to avoid missing fusion determination results due to uncovered combinations.When the RFID-side location status indicates that the tool is in the first operating state, and the video-side tool presence status is valid, the tool is confirmed to be in a normal fusion state. The normal fusion state indicates that both the RFID side and the video side support the tool in a stable state within the permissible range. When the RFID-side location status indicates that the tool is in the second or third operating state, and the video-side tool presence status is uncertain, the tool is confirmed to be in a verification fusion state. The verification fusion state indicates that the RFID side indicates the existence of boundary risks or out-of-bounds risks, but the video side cannot reliably confirm the existence of the tool target, requiring the entry into enhanced sampling or area verification process. When the RFID-side location status is in a signal loss state, and the video-side tool presence status is uncertain, the tool is confirmed to be in an unconfirmed abnormal state. The unconfirmed abnormal state indicates that the RFID side cannot provide a valid location status and the video side cannot reliably confirm the existence of the tool target, requiring the entry into continuous monitoring or manual verification process. The second-stage fusion determination outputs fusion confirmation status, high-risk fusion status, pending fusion status, and abnormal fusion status. If the RFID-side location status and the video-side tool presence status do not directly fall into the above fusion determination results, an intermediate analysis status can be formed during the fusion determination process to assist in judging the status evolution trend or trigger subsequent sampling and verification processes. The above intermediate analysis status is not output as the second-stage fusion determination result.

[0037] In some embodiments of this application, when the graded early warning strategy is activated based on the second-stage fusion determination result, it includes: When the fusion determination result is fusion confirmation, maintain the current monitoring status and do not trigger an alert; When the fusion determination result is a pending fusion status, a low-level early warning strategy is activated to continuously track the corresponding tool and increase the sampling frequency of the RFID side position status and the video side tool presence status. When the fusion determination result is a high-risk fusion state, the intermediate early warning strategy is activated, an abnormal early warning information for the corresponding tool is generated, and the abnormal early warning information is pushed to the management and control platform. When the fusion determination result is an abnormal fusion state, an advanced early warning strategy is activated to generate an emergency early warning message and simultaneously trigger a multi-channel alarm mechanism. Based on the execution results of the tiered early warning strategy, the corresponding early warning event information is recorded to form an early warning processing record.

[0038] Specifically, in the process of initiating a tiered early warning strategy based on the results of the second-stage fusion judgment, the results of the second-stage fusion judgment are first used as the direct basis for early warning decisions. Different fusion judgment results correspond to different early warning handling levels and methods. When the second-stage fusion judgment result is a fusion confirmation state, it means that the RFID-side location status and the video-side tool presence status are consistent within the time correlation window and do not indicate any boundary crossing or abnormal risks. At this time, the current monitoring state is maintained, no early warning is triggered, and the RFID-side location status and the video-side tool presence status are continuously acquired according to the predetermined sampling cycle for subsequent status evolution judgment. When the second-stage fusion judgment result is a pending fusion state, it means that the RFID-side location status indicates that the tool is in the allowed activity area, but the video-side tool presence status cannot be stably confirmed. At this time, a low-level early warning strategy is initiated. The low-level early warning strategy includes entering a continuous tracking mode for the corresponding tool and increasing the sampling frequency of the RFID-side location status and the video-side tool presence status based on the continuous tracking requirements. The increase in sampling frequency is determined according to the security level of the tool control area and the historical frequency of anomalies, in order to acquire status change information in a shorter time interval, thereby completing status confirmation or risk elimination as soon as possible. When the second-stage fusion determination result is a high-risk fusion state, it indicates that the RFID-side location status indicates the tool is in a warning buffer zone or prohibited zone, while the video-side tool presence status is valid. At this time, a medium-level warning strategy is activated. This strategy includes generating abnormal warning information for the corresponding tool. The abnormal warning information must include at least the tool's unique identifier, the area category corresponding to its current location, and the risk type. This abnormal warning information is then pushed to the management platform to prompt timely intervention by management personnel. The triggering conditions and push method for the abnormal warning information are determined based on the risk level rules of the tool control area to ensure the timeliness and controllability of the warning response. When the second-stage fusion determination result is an abnormal fusion state, it indicates that the RFID-side location status is in a signal loss state while the video-side tool presence status is valid. This state is determined to be a high-priority abnormal situation. At this time, an advanced warning strategy is activated. This strategy includes generating emergency warning information and simultaneously triggering a multi-channel alarm mechanism. The multi-channel alarm mechanism includes at least one or more of the following: local audible and visual alarms, management platform alarms, and mobile terminal alarms, to alert to the abnormal situation in the shortest possible time and expand the alarm coverage. The triggering of emergency warning information is directly determined based on the fusion judgment result, without relying on additional threshold judgments, to avoid delayed handling of abnormal states. After the execution of the corresponding graded warning strategy is completed, the fusion judgment result, warning level, warning time, and subsequent handling status of each warning triggering process are recorded to form warning event information. This warning event information is used to construct a complete warning handling record to support subsequent risk retrospective analysis, control effectiveness evaluation, and warning strategy optimization. In another preferred embodiment based on the above embodiments, see [reference needed]. Figure 2 As shown, this embodiment provides a tool positioning and control system based on RFID and video fusion, including: The tool information association module is used to establish an association mapping between the unique code of the RFID tag and the unique identifier of the tool based on the RFID tag on the tool, and to build a tool information file based on the association mapping result. The tool information file includes tool model, usage rights and life management threshold. Configure electronic fence parameters based on the spatial model of the tool control area to determine the boundary range of the permitted activity area, the warning buffer zone and the prohibited area; collect RFID tag signals based on the RFID reader network, perform area attribution estimation based on the RFID tag signals, and determine the area category of the tool based on the area attribution estimation results to generate the RFID side location status. The RFID operation status determination module is used to determine the first stage of the status based on the RFID side location status. When the RFID side location status is in the allowed activity area, it is determined to be the first operation status; when the RFID side location status is in the warning buffer area, it is determined to be the second operation status; when the RFID side location status is in the prohibited area or the RFID tag signal is lost, it is determined to be the third operation status. The video-side knife recognition module is used to synchronously acquire monitoring video streams that are spatially registered with the electronic fence spatial range, process the monitoring video streams, perform knife target detection on the video images, and confirm the knife category and track cross-frame targets of the candidate targets obtained by the knife target detection to obtain the knife presence status on the video side. The fusion status determination module is used to perform a second-stage fusion determination based on the RFID side location status and the video side tool presence status to confirm the tool abnormality type. The strategy execution module is used to activate the corresponding hierarchical early warning strategy according to the anomaly type, and output the early warning information as at least one of local audible and visual alarms, management platform alarms, or mobile terminal alarms.

[0039] Understandably, by combining the RFID-based area attribution sensing capability with the video-based tool target visualization and identification capability, the area attribution status, existence status, and operational risks of tools within the controlled area can be determined synchronously and cross-referenced, thus avoiding misjudgments or omissions caused by relying on a single sensing method. Through the tool information association module, the unique coding of RFID tags and the unique identifier of tools are stably bound, enabling unified management of tool identity information, usage permissions, and lifespan management information, providing a reliable data foundation for subsequent area determination, status identification, and early warning decisions. By introducing an electronic fence-based area division and discrimination mechanism, the activity status of tools can be clearly distinguished into permitted activity areas, early warning buffer areas, and prohibited areas, improving efficiency. The system ensures the controllability and standardization of tool management. A phased status determination structure is constructed using an RFID operation status determination module and a fusion status determination module to effectively identify boundary states, signal anomalies, and inconsistencies in multi-source states, reducing the probability of false or missed warnings. Simultaneously, the continuous detection and tracking of tool targets by the video-side tool identification module enables the system to maintain auxiliary confirmation capabilities even in cases of RFID signal anomalies or unstable area determination. Finally, the strategy execution module triggers warnings in stages based on the fusion determination results, achieving hierarchical and targeted warning responses. This ensures tool management safety while avoiding excessive alarms that could interfere with normal production or workflows, thus improving the overall reliability and practicality of tool positioning control and anomaly warnings.

[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A tool positioning and early warning method based on RFID and video fusion, characterized in that, include: Based on the RFID tags on the cutting tools, an association mapping is established between the unique code of the RFID tag and the unique identifier of the cutting tool, and a cutting tool information file is constructed based on the association mapping result. The cutting tool information file includes the cutting tool model, usage rights and life management threshold. Configure electronic fence parameters based on the spatial model of the tool control area to determine the boundary range of the permitted activity area, the warning buffer zone and the prohibited area; collect RFID tag signals based on the RFID reader network, perform area attribution estimation based on the RFID tag signals, and determine the area category of the tool based on the area attribution estimation results to generate the RFID side location status. The first stage of the operation is determined based on the RFID side location status. When the RFID side location status is in the allowed activity area, it is determined to be the first operating state; when the RFID side location status is in the warning buffer area, it is determined to be the second operating state; when the RFID side location status is in the prohibited area or the RFID tag signal is lost, it is determined to be the third operating state. This is used to synchronously acquire and spatially register the monitoring video stream with the electronic fence spatial range, process the monitoring video stream, perform tool target detection on the video screen, and confirm the tool category and track cross-frame targets of the candidate targets obtained by tool target detection to obtain the tool presence status on the video side. A second-stage fusion determination is performed based on the RFID-side location status and the video-side tool presence status to confirm the tool anomaly type. Based on the type of anomaly, the corresponding hierarchical early warning strategy is activated, and the early warning information is output as at least one of local audible and visual alarms, management platform alarms, or mobile terminal alarms.

2. The tool positioning and early warning method based on RFID and video fusion according to claim 1, characterized in that, When establishing a mapping between the unique code of the RFID tag and the unique identifier of the tool based on the RFID tag on the tool, the following steps are included: Obtain the knives to be included in the management and assign a unique knife identifier to each knife; Read the unique encoding information of the RFID tag attached to the cutting tool, and perform format consistency and repeatability checks on the unique encoding information; Each verified RFID tag's unique code is bound to its corresponding unique tool identifier, forming a unique correspondence. The unique correspondence is stored and marked to obtain the association mapping result between the unique code of the RFID tag and the unique identifier of the tool.

3. The tool positioning and early warning method based on RFID and video fusion according to claim 2, characterized in that, When constructing a tool information archive based on the association mapping results, the following are included: Based on the association mapping result, a corresponding tool file index identifier is generated, and the tool file index identifier is associated with the tool unique identifier; The tool model information, usage permission information, and life management threshold information are obtained based on the tool's unique identifier, and the consistency of the tool model information, usage permission information, and life management threshold information is verified. When the consistency check passes, the model information, usage permission information and life management threshold information are written into the file record corresponding to the tool file index identifier; If the consistency check fails, the corresponding tool file index is marked as abnormal. The tool information file is obtained by integrating the model information, usage permission information and life management threshold information that have passed the consistency verification.

4. The tool positioning and early warning method based on RFID and video fusion according to claim 3, characterized in that, When configuring electronic fence parameters based on the spatial model of the tool control area, the following are included: Acquire spatial structure information of the tool control area, including the area boundary contour, entrance and exit locations, and passable path information; The boundary contour of the tool control area is analyzed, and a closed boundary line is formed according to the boundary contour to obtain the basic fence boundary. Based on the basic fence boundary, the fence extends outward in sequence according to the preset spatial offset rules to form a multi-layered boundary structure of permitted activity area, early warning buffer area and prohibited area. The multi-layer boundary structure is converted into a computable set of electronic fence parameters; The region discrimination rules are determined based on the electronic fence parameter set; The integrity and consistency of the electronic fence parameter set are verified, and the electronic fence parameters are configured when the verification passes, thus obtaining the electronic fence parameters.

5. The tool positioning and early warning method based on RFID and video fusion according to claim 4, characterized in that, When collecting RFID tag signals and generating RFID-side location status based on an RFID reader network, the following are included: Within the tool control area, multiple RFID readers deployed at different spatial locations collect signal data corresponding to RFID tags within the same sampling period. The signal data includes signal strength information and signal continuity information received by each RFID reader. Based on the spatial location and signal strength information of each RFID reader, the RFID tag signal is assigned a region, and the estimated region assignment result of the RFID side is obtained. A consistency analysis is performed on the RFID-side area attribution estimation results obtained within a continuous sampling period. If the area determination results corresponding to the RFID-side area attribution estimation results remain consistent within a continuous sampling period, the RFID-side area attribution estimation results are determined to be valid area estimation results. The effective area estimation results are compared with the permitted activity area, the warning buffer area and the prohibited area to determine the area category where the tool is located; If no valid area estimation result satisfying the consistency condition is obtained within a continuous sampling period, the RFID tag is determined to be in a signal loss state. Generate RFID-side location status based on region category or signal loss status.

6. The tool positioning and early warning method based on RFID and video fusion according to claim 5, characterized in that, When performing the first-stage state determination based on the RFID-side location status, it includes: When the region determination result switches across regions within a continuous sampling period, the tool is determined to be in a boundary determination state. Under the boundary determination state, the corresponding region determination results within the continuous sampling period are summarized and analyzed to determine the region category with the most consecutive occurrences as the boundary determination result. When the boundary determination result corresponds to the allowed activity area, the tool is confirmed to be in the first operating state; When the boundary determination result corresponds to the early warning buffer area, the tool is confirmed to be in the second operating state; When the boundary determination result corresponds to a prohibited area or a signal loss state occurs in the area determination result, the tool is confirmed to be in the third operating state. The confirmed operating status is output as the first-stage status determination result.

7. The tool positioning and early warning method based on RFID and video fusion according to claim 6, characterized in that, When acquiring the tool presence status on the video side, the following is included: Acquire surveillance video streams that are spatially registered with the electronic fence's spatial range, and process the surveillance video streams within a continuous video sampling period; Within each video sampling period, tool target detection is performed on the video frame, and the candidate targets obtained from the tool target detection are confirmed as tool types and tracked across frames to obtain the tool detection results; A consistency analysis is performed on the tool detection results obtained within a continuous video sampling period. If the tool detection results remain consistent within the continuous video sampling period, the tool existence state is determined to be a valid existence state. If no target matching the appearance characteristics of the tool is detected within a continuous video sampling period, or if the tool detection result changes frequently within a continuous video sampling period, the tool's existence state is determined to be an uncertain state. Based on the valid or uncertain state, the video-side tool presence state is generated.

8. The tool positioning and early warning method based on RFID and video fusion according to claim 7, characterized in that, The second-stage fusion determination based on the RFID-side location status and the video-side tool presence status includes: Obtain the RFID-side location status and the video-side tool presence status within the same time-related window, and construct the corresponding status combination; When the RFID-side location status and the video-side tool presence status remain consistent within a continuous time correlation window, the tool is confirmed to be in the corresponding fusion confirmation state. When the RFID-side position status indicates that the tool is in the second or third operating state, and the video-side tool presence status is valid, it is confirmed that the tool is in a high-risk fusion state. When the RFID-side position status indicates that the tool is in the first operating state, and the video-side tool existence status is uncertain, the tool is confirmed to be in the pending confirmation fusion state. When the RFID side position status is signal loss and the video side tool presence status is valid, it is confirmed that the tool is in an abnormal fusion state. The fusion confirmation status, high-risk fusion status, pending confirmation fusion status, or abnormal fusion status are output as the second-stage fusion judgment results.

9. The tool positioning and early warning method based on RFID and video fusion according to claim 8, characterized in that, When the tiered early warning strategy is activated based on the second-stage fusion determination result, it includes: When the fusion determination result is a fusion confirmation state, the current monitoring state is maintained and no warning is triggered; When the fusion determination result is a fusion status to be confirmed, a low-level early warning strategy is activated to continuously track the corresponding tool and increase the sampling frequency of the RFID-side position status and the video-side tool presence status. When the fusion determination result is a high-risk fusion state, an intermediate early warning strategy is activated to generate abnormal early warning information for the corresponding tool and push the abnormal early warning information to the management and control platform. When the fusion determination result is an abnormal fusion state, an advanced early warning strategy is activated, an emergency early warning message is generated, and a multi-channel alarm mechanism is triggered simultaneously. Based on the execution results of the tiered early warning strategy, the corresponding early warning event information is recorded to form an early warning processing record.

10. A tool positioning and control system based on RFID and video fusion, used to implement the tool positioning and early warning method based on RFID and video fusion as described in any one of claims 1-9, characterized in that, include: The tool information association module is used to establish an association mapping between the unique code of the RFID tag and the unique identifier of the tool based on the RFID tag on the tool, and to construct a tool information file based on the association mapping result. The tool information file includes the tool model, usage rights and life management threshold. Configure electronic fence parameters based on the spatial model of the tool control area to determine the boundary range of the permitted activity area, the warning buffer zone and the prohibited area; collect RFID tag signals based on the RFID reader network, perform area attribution estimation based on the RFID tag signals, and determine the area category of the tool based on the area attribution estimation results to generate the RFID side location status. The RFID operation status determination module is used to determine the first stage of the status based on the RFID side location status. When the RFID side location status is in the allowed activity area, it is determined to be in the first operation status; when the RFID side location status is in the warning buffer area, it is determined to be in the second operation status; when the RFID side location status is in the prohibited area or when the RFID tag signal is lost, it is determined to be in the third operation status. The video-side knife recognition module is used to synchronously acquire monitoring video streams that are spatially registered with the electronic fence spatial range, process the monitoring video streams, perform knife target detection on the video images, and confirm the knife category and track cross-frame targets of the candidate targets obtained by the knife target detection to obtain the knife presence status on the video side. The fusion status determination module is used to perform a second-stage fusion determination based on the RFID-side location status and the video-side tool presence status to confirm the tool abnormality type. The strategy execution module is used to activate the corresponding hierarchical early warning strategy according to the anomaly type, and output the early warning information as at least one of local audible and visual alarms, management platform alarms, or mobile terminal alarms.

Images