An intelligent gas cylinder identification and filling matching control system

Abstract

The application discloses an intelligent gas cylinder identification and filling matching control system, and relates to the technical field of gas cylinder filling safety management.The system comprises a gas cylinder identification module, a filling matching module, a visual monitoring module and a data management module.The gas cylinder identification module triggers double-path data collection in response to a gas cylinder arrival detection signal, cross-verify the double-path data with an archive library, and generates a filling permission authentication package.The filling matching module analyzes the permission state and the gas cylinder attribute of the authentication package, and generates a scheduling instruction containing a work station number and control parameters.The visual monitoring module activates an image collector of a target work station, extracts a spatial feature sequence of an operation action, compares the spatial feature sequence with a standard operation feature library to calculate a deviation value, and generates a step completion degree monitoring flow.The data management module aggregates the monitoring flow, the authentication package and the scheduling instruction to form a time sequence record, performs a one-way hash operation on the record data, stores a hash result in association with a previous block, and constructs a tamper-proof traceability chain.The application has the advantages of identity verification, intelligent scheduling, real-time monitoring and reliable traceability.

Description

Technical Field

[0001] This invention relates to the field of gas cylinder filling safety management technology, and in particular to an intelligent gas cylinder identification and filling matching control system. Background Technology

[0002] Gas cylinders, as special equipment for storing and transporting compressed and liquefied gases, are widely used in industrial production, medical and health care, scientific research and teaching, and other fields. Gas cylinder filling operations are highly dangerous; improper operation can lead to serious safety accidents such as explosions and poisoning. Therefore, the gas cylinder filling process requires strict identification, parameter matching, and operational supervision.

[0003] Existing gas cylinder filling management systems suffer from the following technical deficiencies: First, in the identification stage, traditional systems often employ a single identification method, such as relying solely on manual reading of the stamped serial number. A single data source is prone to identification failures due to label detachment or wear, or identity forgery caused by manual label replacement. The lack of cross-verification mechanisms fails to ensure the authenticity and uniqueness of the gas cylinder's identity. Second, regarding filling matching, existing systems lack intelligent analysis capabilities for gas cylinder attributes (such as gas type, pressure rating, and inspection cycle). Workstation allocation relies mainly on manual judgment or a simple first-come, first-served principle, failing to optimize scheduling based on factors such as cylinder type and urgency. This results in low filling efficiency and excessively long waiting times for high-priority cylinders. Finally, in terms of data traceability, existing systems primarily use traditional databases for storing filling records, making the data susceptible to tampering or deletion. In the event of a safety incident, it is difficult to accurately reconstruct the filling process, hindering effective accountability and impeding accident investigation and quality improvement.

[0004] The aforementioned technical deficiencies are particularly prominent in scenarios involving large volumes and diverse types of gas cylinders, affecting filling efficiency and posing safety management risks. There is an urgent need for an integrated control system that combines identity verification, intelligent scheduling, real-time monitoring, and reliable traceability. Summary of the Invention

[0005] This invention provides an intelligent gas cylinder identification and filling matching control system to solve the above-mentioned problems existing in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A smart gas cylinder identification and filling matching control system includes: The gas cylinder identification module is used to trigger dual-channel data acquisition in response to the gas cylinder arrival detection signal, read encrypted identity data from the electronic carrier, extract coded information from the optical carrier, and cross-verify the dual-channel data with the archive to generate a filling license certification package. The filling matching module is used to parse the license status and gas cylinder attributes of the certification package, assign sorting weights based on attribute categories, allocate target workstations in combination with the workstation occupancy matrix, and generate scheduling instructions containing workstation numbers and control parameters. The visual monitoring module is used to activate the image acquisition device of the target workstation according to the scheduling instructions, extract the spatial feature sequence of the operation action, compare it with the standard operation feature library to calculate the deviation value, and generate a step completion monitoring stream. The data management module is used to aggregate monitoring streams, authentication packets, and scheduling instructions to form time-series records, perform one-way hash operations on the recorded data, associate the hash results with the previous blocks for storage, and build a tamper-proof traceability chain.

[0007] Furthermore, the gas cylinder identification module includes: The parallel acquisition submodule is used to simultaneously drive the RFID reader and image acquisition unit after detecting the bit signal to acquire the modulation signal of the electronic tag and the pixel matrix of the QR code. The data extraction submodule is used to demodulate the modulated signal to restore the encrypted byte stream, perform region positioning and encoding restoration on the pixel matrix, and extract the identity serial number and attribute parameter set respectively. The verification generation submodule is used to query historical data from the archive using the identity serial number as an index, compare the relationship between the verification interval and the filling frequency and the threshold, and generate a filling license authentication package.

[0008] Furthermore, the filling and matching module includes: The weight allocation submodule is used to identify the usage category field in the authentication package, assigning the first weight to the medical category, the second weight to the industrial category, and the last weight to the civilian category. The dynamic sorting submodule is used to obtain the entry time of each gas cylinder, calculate the compensation increment corresponding to the waiting time, and add it to the category weight to obtain the comprehensive sorting value. The workstation mapping submodule is used to scan the workstation status table in descending order of sorting value, bind the first gas cylinder in the queue to the first available workstation, look up the table to obtain the corresponding filling parameters, and generate scheduling instructions.

[0009] Furthermore, the visual monitoring module includes: The fixed-point acquisition submodule is used to select the corresponding image sensor according to the workstation number of the scheduling instruction and acquire the image sequence of the operation area according to the preset sampling period; The feature extraction submodule is used to perform differential operations on image sequences to identify motion regions and extract the coordinate change trajectories of human joints and equipment contact points; The deviation calculation submodule is used to measure the similarity between the coordinate trajectory and the feature templates of the standard action library, calculate the execution deviation value of each step, and generate a monitoring stream.

[0010] Furthermore, the parallel acquisition submodule includes: The signal detection unit is used to monitor the gas cylinder entering the predetermined area through a photoelectric sensor array and output a trigger level. The radio frequency activation unit is used to send an interrogation carrier to the electronic tag in response to a trigger level and to receive the reflected identification code signal; An optical capture unit is used to control flash synchronization and shutter action in response to trigger levels, and to record image data of the QR code area.

[0011] Furthermore, the weight allocation submodule includes: The category identification unit is used to parse the category code segment of the authentication packet and determine the gas cylinder service attribute through a lookup table; The numerical mapping unit is used to index the corresponding numerical value in the weight configuration table according to the service attribute, mapping medical attributes to the high-order segment, industrial attributes to the middle-order segment, and civilian attributes to the low-order segment; The weight output unit is used to convert the mapped values ​​into the basic weights for sorting operations.

[0012] Furthermore, the workstation mapping submodule includes: The parameter index unit is used to combine the gas cylinder volume code and the medium type code to form a query key value; The numerical positioning unit uses a key value to perform a binary search in the parameter configuration table to read the pressure limit and flow setpoint. The instruction assembly unit is used to encapsulate the workstation number, gas cylinder identifier, and control values ​​into scheduling instructions according to the protocol format.

[0013] Furthermore, the feature extraction submodule includes: The noise reduction processing unit is used to perform spatial domain smoothing and contrast stretching on the original image; Key point localization unit, used to identify human skeleton nodes and device marker points through a multi-layer feature extraction network; The trajectory generation unit is used to connect the same key points in adjacent frames to form a sequence of motion trajectories.

[0014] Furthermore, the data management module includes: The record integration submodule is used to extract key fields from the three input data sources, and add timestamps and operator codes to form an event record; The hashing submodule is used to convert event records into a bit stream, generate a fixed-length digest through an iterative compression function, and sign the digest with the system key; The chained storage submodule is used to concatenate the signature data with the digest value of the previous data block to form a new data block and establish an index pointer, thus forming a continuous traceability chain.

[0015] Compared with the prior art, the present invention has the following advantages: This invention proposes an intelligent gas cylinder identification and filling matching control system. It simultaneously reads encrypted identity data from an electronic carrier and extracts coded information from an optical carrier, cross-checking these two data streams with a database to form complementary verification. Even if a single carrier fails or is forged, the system can still identify anomalies through data inconsistencies, effectively preventing safety accidents such as incorrect or mixed filling. The system automatically assigns weights based on multi-dimensional attributes of the gas cylinder, such as gas type, pressure level, and urgency, and optimizes allocation based on real-time workstation occupancy, achieving rational utilization of filling resources. By extracting the spatial feature sequence of operational actions and comparing it with a standard feature library in real time, the system can identify and alarm at the instant (millisecond level) of operational deviations, promptly correcting non-standard actions. This proactive monitoring mechanism changes the traditional post-event inspection mode, moving safety control forward to the operation process, significantly reducing safety risks caused by operational errors. By performing a one-way hash operation on each filling record and storing it in association with previous data blocks, a chain-like dependency relationship in time sequence is formed. Any tampering with historical data will result in a hash value mismatch, which will be detected by the system. This technology ensures the integrity and non-repudiation of filling data, providing reliable data support for safety incident investigations, quality traceability, and liability determination.

[0016] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention.

[0017] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a structural diagram of an intelligent gas cylinder identification and filling matching control system according to an embodiment of the present invention; Figure 2 This is a structural diagram of the gas cylinder identification module in an embodiment of the present invention. Detailed Implementation

[0019] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0020] The embodiments of the present invention provide, as follows Figure 1 As shown, an intelligent gas cylinder identification and filling matching control system includes: The gas cylinder identification module is used to trigger dual-channel data acquisition in response to the gas cylinder arrival detection signal, read encrypted identity data from the electronic carrier, extract coded information from the optical carrier, and cross-verify the dual-channel data with the archive to generate a filling license certification package. The filling matching module is used to parse the license status and gas cylinder attributes of the certification package, assign sorting weights based on attribute categories, allocate target workstations in combination with the workstation occupancy matrix, and generate scheduling instructions containing workstation numbers and control parameters. The visual monitoring module is used to activate the image acquisition device of the target workstation according to the scheduling instructions, extract the spatial feature sequence of the operation action, compare it with the standard operation feature library to calculate the deviation value, and generate a step completion monitoring stream. The data management module is used to aggregate monitoring streams, authentication packets, and scheduling instructions to form time-series records, perform one-way hash operations on the recorded data, associate the hash results with the previous blocks for storage, and build a tamper-proof traceability chain.

[0021] The working principle and beneficial effects of the above technical solution are as follows: When a gas cylinder arrives at the filling station, the gas cylinder identification module simultaneously activates the radio frequency reader and image acquisition device to read the encrypted identity information from the electronic tag and extract the encoded attribute parameters from the QR code. The system cross-compares the read dual-channel data with the pre-stored gas cylinder archive to verify the authenticity of the gas cylinder's identity, the validity of the inspection cycle, and the compliance of the filling qualification. After successful verification, an authentication package containing the license status and gas cylinder attributes is generated.

[0022] After receiving the authentication packet, the filling matching module assigns different priority weights based on the gas cylinder's intended use (medical, industrial, or civil), and calculates a comprehensive ranking value based on the cylinder's waiting time. The system monitors the occupancy status of each filling station in real time, allocates available stations to gas cylinders in descending order of ranking value, and queries the parameter database for the corresponding filling pressure, flow rate, and other control parameters of the gas cylinder, encapsulates them into a scheduling command, and sends it to the designated station.

[0023] The visual monitoring module activates the camera at the target workstation according to the scheduling instructions, continuously capturing images of the operator's filling operations. Image processing technology extracts the spatial coordinates of the operator's limb joints and tool / equipment contact points, forming a sequence of motion trajectories. The system compares the actual operation trajectory with pre-stored standard operating procedures, calculates the execution deviation of each step, and generates real-time monitoring stream data to determine operational compliance.

[0024] The data management module aggregates three types of data: authentication packages, scheduling instructions, and monitoring streams, adding timestamps and operator identifiers to form complete event logs. A hash algorithm is executed on the event logs to generate digital digests, which are then digitally signed using the system key. The signed digest value is associated with the digest value of the previous record, forming a chained storage structure to ensure the immutability of historical data and achieve end-to-end traceability.

[0025] In another embodiment, such as Figure 2 As shown, the gas cylinder identification module includes: The parallel acquisition submodule is used to simultaneously drive the RFID reader and image acquisition unit after detecting the bit signal, so as to acquire the modulation signal of the electronic tag and the pixel matrix of the QR code. The data extraction submodule is used to demodulate the modulated signal to restore the encrypted byte stream, perform region positioning and encoding restoration on the pixel matrix, and extract the identity serial number and attribute parameter set respectively. The verification generation submodule is used to retrieve historical data from the archive using the identity serial number as an index, compare the relationship between the inspection interval and filling frequency with the threshold, and generate a filling license certification package. Specifically, the identity serial number is compared with the identification information in the attribute parameter set extracted from the optical carrier to confirm their consistency; the identity serial number is then used as an index to retrieve historical data from the gas cylinder archive database, and the relationship between the inspection interval and filling frequency with the threshold is compared. When the above comparisons are consistent and the inspection and frequency meet the requirements, a filling license certification package is generated.

[0026] The working principle and beneficial effects of the above technical solution are as follows: The parallel acquisition submodule monitors the gas cylinder entering the identification area through a photoelectric sensor array. After the sensor detects an obstruction signal, it outputs a trigger level. This trigger level simultaneously drives the RFID reader and image acquisition unit to start working. The RFID reader transmits a specific frequency interrogation carrier to the electronic tag. After receiving the carrier energy, the electronic tag modulates the stored identity information into the reflected signal and returns it. The image acquisition unit synchronously controls the flash illumination and camera shutter action to capture a high-definition image of the QR code area on the surface of the gas cylinder.

[0027] The data extraction submodule demodulates the modulated signal received by the RFID reader, recovers the encrypted byte stream data, and extracts the unique serial number of the gas cylinder. It also performs region localization on the QR code image acquired by the image acquisition unit, identifies the location markers and coded areas, and uses a decoding algorithm to recover the gas cylinder's volume, medium type, manufacturing date, and other attribute parameters.

[0028] The verification generation submodule uses the identity serial number as the index key to query the historical filling and inspection records of the gas cylinder in the gas cylinder archive. The system reads the last inspection date, calculates the difference between the time interval since then and the specified inspection cycle, and counts the historical filling counts to determine if the rated filling frequency limit has been exceeded. When the inspection interval has not expired and the filling frequency has not exceeded the limit, the system marks the permission status as passed and encapsulates it together with the gas cylinder attribute parameters into an authentication package for output.

[0029] In another embodiment, the filling and matching module includes: The weight allocation submodule is used to identify the usage category field in the authentication package, assigning the first weight to the medical category, the second weight to the industrial category, and the last weight to the civilian category. The dynamic sorting submodule is used to obtain the entry time of each gas cylinder, calculate the compensation increment corresponding to the waiting time, and add it to the category weight to obtain the comprehensive sorting value. The workstation mapping submodule is used to scan the workstation status table (workstation occupancy matrix) in descending order of sorting value, bind the first gas cylinder in the queue to the first available workstation, look up the table to obtain the corresponding filling parameters and generate scheduling instructions.

[0030] The working principle and beneficial effects of the above technical solution are as follows: The weight allocation submodule parses the category code field identifying the service purpose of the gas cylinder in the authentication package and determines the gas cylinder attribute through a preset category mapping table. Medical gas cylinders, due to their highest safety requirements, are assigned a high-level weight (e.g., a value range of 900-999), industrial gas cylinders are assigned a mid-level weight (e.g., a value range of 500-599), and civilian gas cylinders are assigned a low-level weight (e.g., a value range of 100-199). This weight range division ensures that different categories of gas cylinders maintain a clear priority difference in subsequent sorting. Here, high-level, mid-level, and low-level refer to the positional relationship of the numerical ranges: the high-level range represents a larger numerical range and has higher priority in the sorting operation; the mid-level range represents a medium numerical range; and the low-level range represents a smaller numerical range and has the lowest priority.

[0031] The dynamic sorting submodule records the time each gas cylinder enters the waiting queue and calculates the waiting time by the difference between the current time and the entry time. A longer waiting time results in a larger compensation increment. The compensation increment is added to the category's base weight to obtain the comprehensive sorting value. The formula for calculating the comprehensive sorting value is: Where S is the overall ranking value, W is the basic weight of the category, and T is the waiting time. This is a time compensation coefficient. This mechanism ensures both category priority and fairness in waiting.

[0032] The workstation mapping submodule maintains a real-time updated workstation status table, recording the occupancy status of each filling workstation. The system scans the filling queue from highest to lowest based on the comprehensive sorting value, binding the cylinder with the highest sorting value to the first workstation displayed as idle in the status table. A query key is created based on the cylinder's volume specifications and medium type, retrieving the corresponding pressure limit and flow rate setpoint from the filling parameter configuration table. The system assembles the workstation number, cylinder identifier, and control parameters into a scheduling command according to the communication protocol format and sends it to the controller of the corresponding workstation to execute the filling operation.

[0033] In another embodiment, the visual monitoring module includes: The fixed-point acquisition submodule is used to select the corresponding image sensor according to the workstation number of the scheduling instruction and acquire the image sequence of the operation area according to the preset sampling period; The feature extraction submodule is used to perform differential operations on image sequences to identify motion regions and extract the coordinate change trajectories of human joints and equipment contact points; The deviation calculation submodule is used to measure the similarity between the coordinate trajectory and the feature templates of the standard action library, calculate the execution deviation value of each step, and generate a monitoring stream.

[0034] The working principle and beneficial effects of the above technical solution are as follows: After receiving the scheduling instruction, the fixed-point acquisition submodule parses the workstation number field and activates the image sensor corresponding to the workstation through address gating logic. The image sensor continuously captures images of the operation area according to a preset sampling period (e.g., five frames per second), forming an image sequence reflecting the filling operation process.

[0035] The feature extraction submodule first performs noise reduction on the original image sequence, using spatial domain smoothing filtering to eliminate noise interference and contrast stretching to enhance image details. It then performs difference operations on adjacent frames to identify moving regions within the image. Within these moving regions, a multi-layer convolutional feature extraction network is used to locate joint nodes of the human skeleton (such as shoulders, elbows, and wrists) and key equipment markers (such as valve handles and connectors). Finally, it connects the identified key points of the same name in adjacent frames in chronological order to generate a sequence of three-dimensional spatial coordinate trajectories reflecting the operational actions.

[0036] The deviation calculation submodule retrieves the action template corresponding to the current filling type from the standard operation feature library. The action template records the motion trajectory characteristics that should exist at each key point in the standardized operation process. The system calculates the feature similarity between the actually extracted trajectory sequence and the template trajectory, using a dynamic time warping algorithm to measure the degree of matching. For each operation step (such as valve opening, pipeline connection, and pressure adjustment), the trajectory deviation value is calculated separately; a larger deviation value indicates a greater difference between the operation and the standard process. The system summarizes the deviation values, completion times, and execution status of each step to form a monitoring stream data output.

[0037] In another embodiment, the parallel acquisition submodule includes: The signal detection unit is used to monitor the gas cylinder entering the predetermined area through a photoelectric sensor array and output a trigger level. The radio frequency activation unit is used to send an interrogation carrier to the electronic tag in response to a trigger level and to receive the reflected identification code signal. An optical capture unit is used to control flash synchronization and shutter action in response to trigger levels, and to record image data of the QR code area.

[0038] The working principle and beneficial effects of the above technical solution are as follows: Multiple sets of photoelectric sensors are arranged at predetermined positions in the gas cylinder conveying channel, with the sensor transmitters and receivers positioned opposite each other. When a gas cylinder enters the detection area and blocks the light path, the light intensity signal at the receiver changes abruptly. The signal processing circuit performs threshold judgment on the light intensity change, and outputs a high-level trigger signal when the amount of obstruction reaches a set value.

[0039] After receiving a trigger signal, the control logic circuit of the RF activation unit drives the RF transmitting module to generate a carrier signal of a specific frequency. The carrier is radiated to the area where the electronic tag is located through the antenna. The receiving circuit of the electronic tag chip captures the carrier energy and rectifys it into an operating voltage. The internal logic circuit of the chip reads the identification code data from the memory and loads the data into the reflected echo through load modulation. The RF receiving module captures the reflected signal and amplifies, filters, and demodulates it to reconstruct the identification code digital signal.

[0040] Upon receiving the trigger signal, the control circuit of the optical capture unit synchronously sends a flash control pulse and a shutter trigger pulse. The flash unit discharges and emits light instantaneously upon receiving the control pulse, illuminating the QR code area. The camera shutter opens for a specific time under the control of the trigger pulse, exposing the image sensor and converting the light signal into a charge signal, which is then converted from analog to digital to form digital image data.

[0041] In another embodiment, the weight allocation submodule includes: The category identification unit is used to parse the category code segment of the authentication packet and determine the gas cylinder service attribute through a lookup table; The numerical mapping unit is used to index the corresponding numerical value in the weight configuration table according to the service attribute, mapping medical attributes to the high-order segment, industrial attributes to the middle-order segment, and civilian attributes to the low-order segment; The weight output unit is used to convert the mapped values ​​into the basic weights for sorting operations.

[0042] The working principle and beneficial effects of the above technical solution are as follows: The category identification unit locates the category code field from the authentication package data structure. This field uses a fixed position and length encoding. After reading the encoded value, the unit performs a matching search in a pre-stored category lookup table. The lookup table establishes a correspondence between the encoded value and the service attribute (medical, industrial, civil). Upon successful matching, the service attribute identifier of the gas cylinder is output.

[0043] After receiving the service attribute identifier, the numerical mapping unit queries the corresponding numerical range in the weight configuration table using the attribute identifier as the index key. In the configuration table, medical attributes correspond to the high-range numerical range, industrial attributes to the mid-range, and civilian attributes to the low-range. The numerical ranges of different ranges are independent and non-overlapping, and the high-range value is always greater than the mid-range and low-range values. The unit selects a benchmark value from the corresponding range as the category weight for the gas cylinder according to the mapping rules.

[0044] The weight output unit converts the format of the mapped values, adjusting them to the data type and precision required by the sorting module. The output unit is also responsible for weight cache management, storing the weight data of all gas cylinders in the current queue in a high-speed cache for quick retrieval in subsequent sorting operations.

[0045] In another embodiment, the workstation mapping submodule includes: The parameter index unit is used to combine the gas cylinder volume code and the medium type code to form a query key value; The numerical positioning unit uses a key value to perform a binary search in the parameter configuration table to read the pressure limit and flow setpoint. The instruction assembly unit is used to encapsulate the workstation number, gas cylinder identifier, and control values ​​into scheduling instructions according to the protocol format.

[0046] The working principle and beneficial effects of the above technical solution are as follows: The parameter index unit extracts two fields, the cylinder volume code and the medium type code, from the authentication package. The two code values ​​are then concatenated bitwise or joined as strings to form a composite query key. This key-value combination ensures that cylinders of different volumes and media have unique key identifiers.

[0047] The numerical positioning unit uses a binary search algorithm for fast retrieval in the filling parameter configuration table. The configuration table is stored and sorted by key value. The algorithm starts comparing from the middle of the table and recursively narrows the search range based on the key value relationship until a matching item is located. From the matching item, the corresponding filling pressure limit and flow rate setting value for this type of gas cylinder are read. The pressure limit specifies the upper pressure limit at the end of filling, and the flow rate setting value specifies the medium flow rate during the filling process.

[0048] The instruction assembly unit organizes data frames according to the communication protocol format of the control system. The start segment of the data frame is filled with the target station's address number, and the data segments sequentially contain control parameters such as the cylinder identification code, pressure limit, and flow setpoint. A checksum is added to the end of the frame for error detection. The assembled scheduling instruction is sent to the corresponding programmable controller (PLC) via the communication bus. After parsing the instruction parameters, the PLC executes the corresponding filling control action.

[0049] In another embodiment, the feature extraction submodule includes: The noise reduction processing unit is used to perform spatial domain smoothing and contrast stretching on the original image; Key point localization unit, used to identify human skeleton nodes and device marker points through a multi-layer feature extraction network; The trajectory generation unit is used to connect the same key points in adjacent frames to form a sequence of motion trajectories.

[0050] The working principle and beneficial effects of the above technical solution are as follows: The noise reduction processing unit preprocesses the original image sequence frame by frame. Spatial domain smoothing uses median filtering or Gaussian filtering algorithms to suppress random noise through pixel neighborhood operations. Contrast stretching expands the effective grayscale range of the image through a grayscale mapping function, enhancing the distinction between the target and the background and facilitating subsequent feature detection.

[0051] The keypoint localization unit employs a deep convolutional neural network for feature extraction. The network takes a pre-processed image as input and extracts features from edges and textures to higher-level semantics through multiple layers of convolution, pooling, and non-linear activation operations. The network's output layer predicts the coordinates of key nodes in the human skeleton and device markers within the image. Through pre-supervised training, the network learns the mapping relationship between features and locations, enabling accurate identification of keypoint locations under different poses in practical applications.

[0052] The trajectory generation unit maintains a keypoint tracking table, recording the coordinate sequence of each keypoint in historical frames. When keypoint localization is completed in a new frame, the unit matches and associates the new detected point with historical points based on the principle of closest spatial distance. The coordinates of the same keypoint in adjacent frames are connected in chronological order to form the point's motion trajectory. The trajectory data is stored as a three-dimensional coordinate sequence, containing both temporal and spatial dimensions, fully reflecting the changes in motion during the operation.

[0053] In another embodiment, the data management module includes: The record integration submodule is used to extract key fields from the three input data sources, and add timestamps and operator codes to form an event record; The hashing submodule is used to convert event records into a bit stream, generate a fixed-length digest through an iterative compression function, and sign the digest with the system key; The chained storage submodule is used to concatenate the signature data with the digest value of the previous data block to form a new data block and establish an index pointer, thus forming a continuous traceability chain.

[0054] The working principle and beneficial effects of the above technical solution are as follows: The recording and integration submodule sets up a data buffer to receive three data streams in real time: the authentication package from the gas cylinder identification module, the scheduling instructions from the filling and matching module, and the monitoring stream from the visual monitoring module. The unit extracts core fields from each data stream, such as the gas cylinder identifier and attributes in the authentication package, the workstation number and control parameters in the scheduling instructions, and the operation deviation and completion status in the monitoring stream. The system reads the current timestamp and operator login identifier, and assembles the extracted fields, timestamp, and operator code into a complete event record according to a predetermined data structure.

[0055] The hashing submodule serializes event records into a continuous bit stream, which serves as the input to the hash function. The hash function employs an iterative compression structure, compressing input data of arbitrary length into a fixed-length digest value through multiple rounds of mixed operations. The digest value is unidirectional and collision-resistant; even a small change in the original data will result in a completely different digest value. The system uses a preset key to perform digital signature operations on the digest value. The signature result can only be verified by the system holding the key, ensuring the trustworthiness of the data source.

[0056] The linked storage submodule reads the digest value of the previous data block from memory and concatenates it with the current signature data to form a new data block. Each new data block is assigned a unique block identifier, and an index pointer pointing to the previous block is established, forming a linked structure. Each data block contains three parts: the current event record, the current digest signature, and the digest value of the previous block. Because the digest of each block depends on the content of the previous block, any tampering with historical data will cause the digest verification of all subsequent blocks to fail, thus achieving data tamper-proofing and end-to-end traceability.

[0057] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from the spirit and scope of the invention.

Claims

1. An intelligent gas cylinder identification and filling matching control system, characterized in that, include: The gas cylinder identification module is used to trigger dual-channel data acquisition in response to the gas cylinder arrival detection signal, read encrypted identity data from the electronic carrier, extract coded information from the optical carrier, and cross-verify the dual-channel data with the archive to generate a filling license certification package. The gas cylinder identification module includes: The parallel acquisition submodule is used to simultaneously drive the RFID reader and image acquisition unit after detecting the bit signal, to acquire the modulation signal of the electronic tag and the pixel matrix of the QR code; the parallel acquisition submodule includes: The signal detection unit is used to monitor the gas cylinder entering the predetermined area through a photoelectric sensor array and output a trigger level. The radio frequency activation unit is used to send an interrogation carrier to the electronic tag in response to a trigger level and to receive the reflected identification code signal; The optical capture unit is used to control flash synchronization and shutter action in response to trigger levels, and to record image data of the QR code area; The data extraction submodule is used to demodulate the modulated signal to restore the encrypted byte stream, perform region positioning and encoding restoration on the pixel matrix, and extract the identity serial number and attribute parameter set respectively. The verification generation submodule is used to query the archive to obtain historical data using the identity serial number as an index, compare the relationship between the verification interval and the filling frequency and the threshold, and generate a filling license authentication package. The filling matching module is used to parse the license status and cylinder attributes of the certification package, assign sorting weights based on attribute categories, allocate target workstations in conjunction with the workstation occupancy matrix, and generate scheduling instructions containing workstation numbers and control parameters. The filling matching module includes: The weight allocation submodule is used to identify the usage category field in the authentication package, assigning a first weight to medical-use categories, a second weight to industrial-use categories, and a last weight to civilian-use categories. The weight allocation submodule includes: The category identification unit is used to parse the category code segment of the authentication packet and determine the gas cylinder service attribute through a lookup table; The numerical mapping unit is used to index the corresponding numerical value in the weight configuration table according to the service attribute, mapping medical attributes to the high-order segment, industrial attributes to the middle-order segment, and civilian attributes to the low-order segment; The weight output unit is used to convert the mapped values ​​into the basic weights for sorting operations; The dynamic sorting submodule is used to obtain the entry time of each gas cylinder, calculate the compensation increment corresponding to the waiting time, and add it to the category weight to obtain the comprehensive sorting value. The workstation mapping submodule is used to scan the workstation status table in descending order of sorting value, bind the first gas cylinder in the queue to the first available workstation, look up the table to obtain the corresponding filling parameters and generate scheduling instructions. The visual monitoring module is used to activate the image acquisition device of the target workstation according to the scheduling instructions, extract the spatial feature sequence of the operation action, compare it with the standard operation feature library to calculate the deviation value, and generate a step completion monitoring stream. The data management module is used to aggregate monitoring streams, authentication packets, and scheduling instructions to form time-series records, perform one-way hash operations on the recorded data, associate the hash results with the previous blocks for storage, and build a tamper-proof traceability chain.

2. The intelligent gas cylinder identification and filling matching control system according to claim 1, characterized in that, The visual monitoring module includes: The fixed-point acquisition submodule is used to select the corresponding image sensor according to the workstation number of the scheduling instruction and acquire the image sequence of the operation area according to the preset sampling period; The feature extraction submodule is used to perform differential operations on image sequences to identify motion regions and extract the coordinate change trajectories of human joints and equipment contact points; The deviation calculation submodule is used to measure the similarity between the coordinate trajectory and the feature templates of the standard action library, calculate the execution deviation value of each step, and generate a monitoring stream.

3. The intelligent gas cylinder identification and filling matching control system according to claim 1, characterized in that, The workstation mapping submodule includes: The parameter index unit is used to combine the gas cylinder volume code and the medium type code to form a query key value; The numerical positioning unit uses a key value to perform a binary search in the parameter configuration table to read the pressure limit and flow setpoint. The instruction assembly unit is used to encapsulate the workstation number, gas cylinder identifier, and control values ​​into scheduling instructions according to the protocol format.

4. The intelligent gas cylinder identification and filling matching control system according to claim 2, characterized in that, The feature extraction submodule includes: The noise reduction processing unit is used to perform spatial domain smoothing and contrast stretching on the original image; Key point localization unit, used to identify human skeleton nodes and device marker points through a multi-layer feature extraction network; The trajectory generation unit is used to connect the same key points in adjacent frames to form a sequence of motion trajectories.

5. The intelligent gas cylinder identification and filling matching control system according to claim 1, characterized in that, The data management module includes: The record integration submodule is used to extract key fields from the three input data sources, and add timestamps and operator codes to form an event record; The hashing submodule is used to convert event records into a bit stream, generate a fixed-length digest through an iterative compression function, and sign the digest with the system key; The chained storage submodule is used to concatenate the signature data with the digest value of the previous data block to form a new data block and establish an index pointer, thus forming a continuous traceability chain.

Images