Reinforcement cage identification code tracking method and system

Abstract

The invention relates to a reinforcement cage identification code tracking method and system.The method comprises the steps that after a reinforcement cage is manufactured, identification code jet printing operation is executed according to an identification code, and after jet printing is completed, the identification code is bound with a preset RFID tag on the reinforcement cage; after jet printing of the reinforcement cage is completed, an identification code image on the reinforcement cage is collected, image preprocessing is conducted on the identification code image, and the preprocessed identification code image is obtained; evaluating the preprocessed identification code image through a pre-trained evaluation model to obtain an identification confidence score; in the circulation process of the reinforcement cage, based on the RFID tag bound with the identification code, data of the RFID tag are collected in real time through a multi-source read-write device, and movement track information of the reinforcement cage is constructed based on the data of the RFID tag; and judging whether a path is abnormal or not based on the moving track information. The method has the effect of improving the tracking efficiency of the identification code of the reinforcement cage.

Description

Technical Field

[0001] This application relates to the technical field of construction information management, and in particular to a method and system for tracking steel cage identification codes. Background Technology

[0002] Currently, in large-scale engineering construction sites, rebar cages, as core structural components, are often prefabricated by multiple processing workshops and then transported to designated construction units for installation. During this process, to achieve full lifecycle traceability management of the components, paper documents or simple barcode systems are used to record and track the production process and on-site information of the rebar cages.

[0003] Existing methods for marking steel cages mainly rely on static information identification, such as spraying two-dimensional barcodes on the components and recording them into the system through manual scanning or automatic identification, thereby tracking their location and status.

[0004] The existing technical solutions mentioned above have the following drawbacks: the existing identification code generation and printing process lacks an evaluation mechanism for image recognizability. Often, the printing quality is found to be substandard only when subsequent recognition fails or data is missing, resulting in the components being unable to be effectively identified and tracked, thus affecting the construction progress. Therefore, there is room for improvement. Summary of the Invention

[0005] To improve the tracking efficiency of rebar cage identification codes, this application provides a rebar cage identification code tracking method and system.

[0006] The above-mentioned objective of this application is achieved through the following technical solution: A method for tracking steel cage identification codes, the method comprising: Obtain the steel cage production order information and generate a corresponding identification code based on the steel cage production order information; After the steel cage is manufactured, the identification code is printed according to the identification code. After the printing is completed, the identification code is bound to the RFID tag preset on the steel cage for use as a unique identifier for subsequent tracking and recording. After the printing of the steel cage is completed, the image of the identification code on the steel cage is captured, and the image of the identification code is preprocessed to obtain the preprocessed identification code image. The preprocessed identification code image is evaluated using a pre-trained evaluation model to obtain a recognition confidence score; The identification confidence score is compared with a preset identification threshold. If the identification confidence score is lower than the preset identification threshold, the subsequent steel cage transfer steps are paused and a reprinting operation is performed until the identification confidence score reaches the preset identification threshold. During the transfer of the steel cage, based on the RFID tag bound to the identification code, the data of the RFID tag is collected in real time using a multi-source reader / writer device, and the movement trajectory information of the steel cage is constructed based on the data of the RFID tag. Based on the movement trajectory information, it is determined whether there is any path abnormality. When abnormal behavior is determined, an abnormality marker will be triggered, and the anti-counterfeiting information will be bound to the tracking record associated with the identification code.

[0007] By adopting the above technical solutions, and generating corresponding identification codes based on the steel cage production order information, each steel cage can have a unique identification mark from the initial stage of production, thereby ensuring accurate tracking and traceability of the steel cage in subsequent stages. After the steel cage is manufactured, the identification code is printed and bound to a pre-set RFID tag, enabling bidirectional association between the physical steel cage and electronic data, thus enhancing the consistency and integrity of tracking records. By acquiring identification code images and performing image preprocessing, the quality and stability of images during subsequent identification processes can be improved, thereby increasing identification accuracy and reducing decoding failure rate. By generating an identification confidence score through an evaluation model and comparing it with an identification threshold to determine whether to perform reprinting, the quality of the identification can be dynamically guaranteed, thereby improving the reliability of the overall identification system. By constructing movement trajectory information based on RFID and judging path anomalies, real-time monitoring of the steel cage circulation process can be achieved, thereby promptly identifying abnormal operating behaviors and preventing safety risks caused by forgery or misoperation.

[0008] In one example, this application can be further configured as follows: generating a corresponding identification code based on the rebar cage production order information specifically includes: Extract the production information of the steel cage from the steel cage production order information. The production information includes component number, production batch, mold number, and target construction unit information. The production information is combined according to a preset encoding rule to generate a two-dimensional display code with embedded semantic fields, which serves as the identification code to uniquely identify the steel cage.

[0009] By adopting the above technical solution, by extracting information such as component number, production batch, mold number and construction unit from the steel cage production order information, and generating a two-dimensional display code with embedded semantic fields based on preset coding rules, the generated identification code can not only be unique, but also carry multi-dimensional production semantic information, thereby improving the information transmission efficiency and interpretation value of the identification code in the production, transportation and construction stages.

[0010] In one example, this application can be further configured such that: the step of performing the identification code printing operation based on the identification code specifically includes: Based on the identification code, initial printing parameters are generated, including printing power, nozzle pressure and character spacing. The surface environment parameters of the reinforcing cage are collected, and the initial printing parameters are dynamically adjusted based on the surface environment parameters to obtain the final printing parameters. The identification code printing operation is then performed according to the final printing parameters. The surface environment parameters include temperature and humidity, material type, and surface roughness.

[0011] By adopting the above technical solution, the initial printing parameters are generated based on the identification code, and the printing parameters are dynamically adjusted in combination with the collected surface environmental parameters of the steel cage. This enables the printing operation to adapt to different materials, temperature, humidity and roughness conditions, thereby ensuring that the printed content has good clarity and adhesion in various environments, and improving the durability and recognizability of the markings.

[0012] In one example, this application can be further configured such that: the image preprocessing of the identification code image to obtain a preprocessed identification code image specifically includes: The identification code image is subjected to image grayscale conversion, noise filtering and edge sharpening operations to enhance the clarity and contour contrast of the identification code image, resulting in an enhanced identification code image. The enhanced identification code image is subjected to size normalization, angle correction, and distortion correction operations to ensure that the image has a uniform recognition structure under different acquisition conditions, and finally the preprocessed identification code image is obtained.

[0013] By adopting the above technical solutions, operations such as grayscale conversion, noise filtering, and edge sharpening are performed on the identification code image to enhance the detail and contrast of the image, thereby improving the prominence of the identification area in the image. By performing size normalization, angle correction, and distortion correction on the enhanced image, the image structure can be unified to adapt to the input requirements of the subsequent evaluation model, thereby improving the stability and applicability of the recognition model under complex acquisition conditions.

[0014] In one example, this application can be further configured such that the rebar cage identification code tracking method also includes: Collect a dataset of identification code image samples and manually annotate the dataset. The annotation labels are used to indicate whether the identification code image can be successfully recognized during the actual decoding process. The dataset of identification code image samples includes recognizable image samples and unrecognizable image samples. An image recognition rate evaluation model is constructed using a convolutional neural network structure, and the image recognition rate evaluation model is trained in a supervised manner using the labeled tags as supervision signals to obtain the pre-trained image recognition rate evaluation model.

[0015] By adopting the above technical solution, collecting identification code image samples and manually annotating them, and using identifiable and unidentifiable labels as supervisory signals to train the image evaluation model, the model can accurately judge the identifiability of the identification image. In practical applications, this allows for dynamic detection of whether the printing effect meets the usage requirements, ensuring that the identification code always has effective identification capabilities throughout its entire lifecycle.

[0016] In one example, this application can be further configured such that, before comparing the identification confidence score with a preset identification threshold, the rebar cage identification code tracking method further includes: The environmental state parameters of the identification code image are collected, and the preset recognition threshold is dynamically adjusted based on the environmental state parameters. The environmental state parameters include the average image brightness, image clarity index, and occlusion degree index.

[0017] By adopting the above technical solution, and by collecting environmental state parameters such as average image brightness, sharpness index, and occlusion degree, and dynamically adjusting the recognition threshold accordingly, misjudgment caused by environmental changes can be avoided, thereby improving the robustness of the evaluation mechanism, making the confidence comparison more in line with actual recognition conditions, and enhancing the adaptability and reliability of image evaluation.

[0018] In one example, this application can be further configured such that: determining whether there is a path anomaly based on the movement trajectory information specifically includes: Based on the target construction unit information corresponding to the identification code, a target path template is generated by matching from the preset path rules. The target path template is used to indicate the standard transfer path of the steel cage. The movement trajectory information is matched with the target path template for similarity analysis. If the similarity is lower than the anomaly judgment threshold, or if there are abnormal node events in the movement trajectory information, it is determined that the current rebar cage has a path anomaly. The abnormal node events include path interruption, frequent backtracking, or exceeding the boundary of the construction area.

[0019] By adopting the above technical solution, and by generating a target path template based on the target construction unit information and performing similarity matching analysis with the movement trajectory information, it is possible to accurately identify whether the transfer path deviates from the planned route; by identifying abnormal node events such as path interruption, frequent backtracking or boundary crossing, it is possible to detect violations in the process of steel cage transportation, thereby triggering abnormal markers and binding anti-counterfeiting information in a timely manner, and strengthening quality and safety control and anti-tampering capabilities.

[0020] The second objective of this invention is achieved through the following technical solution: A rebar cage identification code tracking system, the rebar cage identification code tracking system comprising: The identification code generation module is used to obtain steel cage production order information and generate a corresponding identification code based on the steel cage production order information. The identification code printing and binding module is used to perform identification code printing operation according to the identification code after the steel cage is manufactured, and bind the identification code to the preset RFID tag on the steel cage after printing, which is used as a unique identifier for subsequent tracking and recording. The image preprocessing module is used to acquire the identification code image on the steel cage after the steel cage is printed, and to perform image preprocessing on the identification code image to obtain the preprocessed identification code image. The image evaluation module is used to evaluate the preprocessed identification code image using a pre-trained evaluation model to obtain a recognition confidence score. The confidence comparison module is used to compare the recognition confidence score with a preset recognition threshold. If the recognition confidence score is lower than the preset recognition threshold, the subsequent steel cage transfer steps are paused and a reprinting operation is performed until the recognition confidence score reaches the preset recognition threshold. The trajectory acquisition module is used to collect data from RFID tags bound to the identification code in real time during the transfer of the steel cage using a multi-source reader / writer device, and to construct the movement trajectory information of the steel cage based on the data from the RFID tags. The path anomaly detection module is used to determine whether there is a path anomaly based on the movement trajectory information. When an abnormal behavior is detected, an anomaly marker will be triggered, and the anti-counterfeiting information will be bound to the tracking record associated with the identification code.

[0021] By adopting the above technical solutions, and generating corresponding identification codes based on the steel cage production order information, each steel cage can have a unique identification mark from the initial stage of production, thereby ensuring accurate tracking and traceability of the steel cage in subsequent stages. After the steel cage is manufactured, the identification code is printed and bound to a pre-set RFID tag, enabling bidirectional association between the physical steel cage and electronic data, thus enhancing the consistency and integrity of tracking records. By acquiring identification code images and performing image preprocessing, the quality and stability of images during subsequent identification processes can be improved, thereby increasing identification accuracy and reducing decoding failure rate. By generating an identification confidence score through an evaluation model and comparing it with an identification threshold to determine whether to perform reprinting, the quality of the identification can be dynamically guaranteed, thereby improving the reliability of the overall identification system. By constructing movement trajectory information based on RFID and judging path anomalies, real-time monitoring of the steel cage circulation process can be achieved, thereby promptly identifying abnormal operating behaviors and preventing safety risks caused by forgery or misoperation.

[0022] In summary, this application includes the following beneficial technical effects: 1. By generating corresponding identification codes based on the steel cage production order information, each steel cage can have a unique identification mark at the beginning of production, thereby ensuring accurate tracking and traceability of the steel cage in subsequent stages; by performing identification code printing and binding a preset RFID tag after the steel cage is manufactured, a two-way association between the physical steel cage and electronic data can be achieved, thereby enhancing the consistency and integrity of tracking records. 2. By acquiring and preprocessing the identification code images, the quality and stability of the images can be improved in the subsequent recognition process, thereby increasing the recognition accuracy and reducing the decoding failure rate. By generating a recognition confidence score through the evaluation model and comparing it with the recognition threshold to determine whether to perform reprinting, the quality of the identification can be dynamically guaranteed, thereby improving the reliability of the overall identification system. By constructing movement trajectory information based on RFID and judging path anomalies, real-time monitoring of the steel cage transfer process can be achieved, thereby timely identifying abnormal operation behavior and preventing safety risks caused by counterfeiting or misoperation. Attached Figure Description

[0023] Figure 1 This is a flowchart of a steel cage identification code tracking method according to an embodiment of this application; Figure 2 This is a flowchart illustrating the implementation of step S10 in a steel cage identification code tracking method according to an embodiment of this application. Figure 3 This is a flowchart illustrating the implementation of step S20 in a steel cage identification code tracking method according to an embodiment of this application. Figure 4 This is a flowchart illustrating the implementation of step S30 in a steel cage identification code tracking method according to an embodiment of this application. Figure 5 This is a flowchart of an implementation of a steel cage identification code tracking method in one embodiment of this application; Figure 6 This is another implementation flowchart of a steel cage identification code tracking method in one embodiment of this application; Figure 7 This is a flowchart illustrating the implementation of step S70 in a steel cage identification code tracking method according to an embodiment of this application. Figure 8 This is a principle block diagram of a steel cage identification code tracking system according to one embodiment of this application. Detailed Implementation

[0024] The present application will be further described in detail below with reference to the accompanying drawings.

[0025] In one embodiment, such as Figure 1 As shown, this application discloses a method for tracking steel cage identification codes, which specifically includes the following steps: S10: Obtain the steel cage production order information and generate the corresponding identification code based on the steel cage production order information.

[0026] Specifically, the extracted steel cage production order information includes, but is not limited to, project number, component number, work group information, and planned delivery time. By calling the coding rule module, the above fields are combined in a preset order to generate a standardized identification string, which is then converted into a structurally stable QR code pattern format. During the identification code generation process, a timestamp and anti-duplicate mark are also added to ensure the uniqueness and timeliness of the coded content. For example, in the production task of "Project ZJ-20250823" with component number "SG-0087", the generated identification code is "ZJ20250823-SG0087-BZ05-T143200", which is encapsulated in QR code form for subsequent processing.

[0027] S20: After the steel cage is manufactured, the identification code is printed according to the identification code. After printing, the identification code is bound to the RFID tag preset on the steel cage for use as a unique identifier for subsequent tracking and recording.

[0028] Specifically, after the rebar cage manufacturing process is completed, the printing control command is automatically invoked to send the corresponding identification code to the inkjet printer control terminal. The printing equipment processes the pixel arrangement of the QR code according to the printing template and completes the printing operation on the designated area of ​​the rebar cage surface. Then, it reads the RFID tag information that has been preset on the surface of the rebar cage and associates the identification code content with the unique code of the RFID tag through the data writing logic, thereby establishing a unique identification link that can be used for subsequent tracking. For example, a finished rebar cage with a diameter of 12mm and a length of 6m has a QR code "ZJ20250823-SG0087-BZ05-T143200" printed at the discharge port and is bound to its RFID tag number "RFID_6A9E7C2D".

[0029] S30: After the printing of the steel cage is completed, the image of the identification code on the steel cage is acquired, and the image of the identification code is preprocessed to obtain the preprocessed identification code image.

[0030] Specifically, a high-definition industrial camera is activated to capture images of the printing area. The captured images are first processed to grayscale to enhance the clarity of the QR code outline edges. Then, a bilateral filtering algorithm is used to denoise the images to eliminate external interference factors such as oil stains and water spots. At the same time, image rotation correction and size standardization are performed to ensure that the QR code image meets the input specifications of the recognition model, thereby generating a pre-processed image with a complete structure. For example, if there are light spot interference and shadow occlusion in the original image, a clear corrected image with a resolution of 640×640 is generated after processing for subsequent recognition.

[0031] S40: The pre-processed identification code image is evaluated using a pre-trained evaluation model to obtain a recognition confidence score.

[0032] Specifically, the preprocessed identification code image is fed into a pre-trained evaluation model. This evaluation model is built on a convolutional neural network architecture, and its internal parameters have been trained using a large number of printed QR code samples. The evaluation model will output a recognition confidence score based on the multi-dimensional feature extraction results of the image, such as pixel coherence, edge sharpness, and contrast balance. This score is presented in decimal form. The higher the value, the higher the recognition level of the QR code image. For example, if the recognition confidence of a certain printed image is 0.92, it means that the image quality meets the requirements for subsequent scanning and reading under the current shooting conditions.

[0033] S50: Compare the recognition confidence score with the preset recognition threshold. If the recognition confidence score is lower than the preset recognition threshold, pause the subsequent steel cage transfer steps and perform a reprinting operation until the recognition confidence score reaches the preset recognition threshold.

[0034] Specifically, after obtaining the recognition confidence score, the recognition judgment logic is executed. The current score is compared with the recognition threshold set in the local configuration file. When the judgment result is that the threshold requirement is not met, the current rebar cage is marked as pending re-spraying, and the subsequent transfer process of the rebar cage on the current production line is paused. At the same time, the printing step is called back and the collection and evaluation process is repeated until the recognition confidence score meets the threshold requirement. Then, its status is updated to the transferable status and the transfer process is resumed. For example, when the confidence of a rebar cage print image is only 0.68 and the threshold is set to 0.85, the re-spraying process will be triggered and it will be prohibited from entering the next storage area.

[0035] S60: During the transfer of the steel cage, based on the RFID tags bound to the identification code, multi-source reading and writing devices are used to collect the data of the RFID tags in real time, and the movement trajectory information of the steel cage is constructed based on the data of the RFID tags.

[0036] Specifically, after the rebar cage enters the circulation process, the reading and writing device interface program is invoked to control multiple RFID reading and writing nodes deployed at the production line exit, transportation handover point, and construction site entrance to poll and scan the tag data in the area. After reading the bound unique RFID number, the timestamp and device number information are recorded, and the data is transmitted to the trajectory recording module and added to the circulation trajectory sequence of the rebar cage in turn. Combined with the geographical location information of each node, a time-series trajectory path is established, thereby realizing the visualized path tracking of the entire transportation and construction process of the rebar cage. For example, in the three stages of "processing point → storage yard → tower crane platform", the RFID number of the same component can be continuously tracked through different nodes and the dwell time.

[0037] S70: Based on the movement trajectory information, determine whether there is a path anomaly. When an abnormal behavior is determined, an anomaly marker will be triggered, and the anti-counterfeiting information will be bound to the tracking record associated with the identification code.

[0038] Specifically, the trajectory analysis logic compares the consistency of the rebar cage's flow path. If two or more consecutive nodes jump, there are illegal nodes that do not exist in the preset process flow path, or nodes that have not been updated for a long time, it is determined to be abnormal path behavior. The abnormal state marking logic is triggered to mark the record as suspicious. At the same time, the anti-counterfeiting information writing instruction is called to add an abnormal identifier field and time node information to the identifier tracking record for subsequent traceability verification. For example, if a rebar cage should have entered the "quality inspection area" from the "welding area", but its next record appears in the "loading area", it can be determined to be an abnormal path and an abnormal identifier "PATH_ERR#PRSKIP@T165230" is added for anti-counterfeiting review.

[0039] By adopting the above technical solutions, and generating corresponding identification codes based on the steel cage production order information, each steel cage can have a unique identification mark from the initial stage of production, thereby ensuring accurate tracking and traceability of the steel cage in subsequent stages. After the steel cage is manufactured, the identification code is printed and bound to a pre-set RFID tag, enabling bidirectional association between the physical steel cage and electronic data, thus enhancing the consistency and integrity of tracking records. By acquiring identification code images and performing image preprocessing, the quality and stability of images during subsequent identification processes can be improved, thereby increasing identification accuracy and reducing decoding failure rate. By generating an identification confidence score through an evaluation model and comparing it with an identification threshold to determine whether to perform reprinting, the quality of the identification can be dynamically guaranteed, thereby improving the reliability of the overall identification system. By constructing movement trajectory information based on RFID and judging path anomalies, real-time monitoring of the steel cage circulation process can be achieved, thereby promptly identifying abnormal operating behaviors and preventing safety risks caused by forgery or misoperation.

[0040] In one embodiment, such as Figure 2 As shown, in step S10, the corresponding identification code is generated based on the steel cage production order information, specifically including: S11: Extract the production information of the steel cage from the steel cage production order information. The production information includes component number, production batch, mold number and target construction unit information.

[0041] Specifically, the component number, production batch, mold number, and target construction unit information fields are sequentially read from the received rebar cage production order data structure. The component number indicates the type of structural component to which the rebar cage belongs; the production batch identifies the production task batch to which the rebar cage belongs; the mold number locates the mold template used to make the rebar cage; and the target construction unit information indicates the floor or component location where the rebar cage will be installed. The above information is extracted from the original order data in JSON or XML format using structured parsing, and field standardization and encoding conversion are performed to ensure consistency in subsequent operations. For example, if a certain order data records a component number of "CJK-0102", a production batch of "B2308A", a mold number of "MDA-06", and a target construction unit of "East side of the fourth floor of Tower B", then the extracted production information is a combination structure of the four field values, which is used to generate identification codes and establish identity mappings.

[0042] S12: Combine production information according to preset coding rules to generate a two-dimensional display code with embedded semantic fields, which serves as an identification code to uniquely identify the steel cage.

[0043] Specifically, the component number, production batch, mold number, and target construction unit information are concatenated and encoded according to a preset coding structure. The coding rules include field order, field length standards, separator settings, and validation field generation logic. After concatenation, an original coding string that meets the input format requirements is formed. Then, a QR code encoding algorithm is used to generate a two-dimensional display code with error correction capabilities from this coding string. During the encoding process, key fields are embedded in the main code area of ​​the QR code in a readable structure for subsequent rapid identification. Finally, a structurally unique and semantically clear two-dimensional identifier image is obtained as the unique identifier of the rebar cage. For example, for the four items "CJK-0102", "B2308A", "MDA-06" and "B-4-E", after the coding rules are converted, a structured coding string "CJK0102#B2308A#MDA06#B4E" is generated, and a corresponding QR code image is generated for printing binding and full-process tracking.

[0044] In one embodiment, such as Figure 3 As shown, in step S20, i.e., according to the identification code, the identification code printing operation is performed, which specifically includes: S21: Generate initial printing parameters based on the identification code. The printing parameters include printing power, nozzle pressure, and character spacing.

[0045] Specifically, after receiving the identification code content to be printed, the corresponding printing parameter template is extracted from the preset parameter mapping table according to the code length and character complexity. The initial printing power, nozzle pressure and character spacing values ​​are generated by combining the current device model and ink type. The printing power is used to control the ink jet intensity, the nozzle pressure affects the clarity of ink droplet formation, and the character spacing determines whether the code structure is clear and recognizable. For example, for a QR code containing 25 characters, the matching rule sets the corresponding initial printing power to 80%, the nozzle pressure to 0.3MPa, and the character spacing to 0.8mm. The generated initial parameters serve as the basic input values ​​for the printing control algorithm to ensure that the printing process has basic readability and structural integrity.

[0046] S22: Collect surface environmental parameters of the rebar cage, dynamically adjust the initial printing parameters based on the surface environmental parameters to obtain the final printing parameters, and perform the identification code printing operation according to the final printing parameters. The surface environmental parameters include temperature and humidity, material type and surface roughness.

[0047] Specifically, the surface environment detection interface is invoked to collect real-time information on temperature and humidity, material type, and surface roughness indices of the current rebar cage printing area. Temperature and humidity are obtained through an infrared temperature and humidity sensor, material type is determined by identifying the steel code or a preset parameter table, and roughness is calculated based on image sampling and grayscale gradient analysis. These parameters are then dynamically corrected in conjunction with the initial printing parameters. For example, when the rebar cage material is detected to be HRB400, the surface temperature of the printing area is 45℃, the relative humidity is 70%, and the roughness reaches 2.5μm, the control logic increases the printing power to 90%, the nozzle pressure to 0.35MPa, and fine-tunes the character spacing to 1.0mm to improve printing adhesion and image clarity. Finally, the printing action is initiated based on the corrected printing parameters to complete the imaging output of the identification code image on the surface of the rebar cage.

[0048] In one embodiment, such as Figure 4 As shown, in step S30, the identification code image is preprocessed to obtain a preprocessed identification code image, specifically including: S31: Perform image grayscale conversion, noise filtering, and edge sharpening operations on the identification code image to enhance the clarity and contour contrast of the identification code image, resulting in an enhanced identification code image.

[0049] Specifically, by loading the acquired identification code image data, the image grayscale conversion operation is first performed to convert the color image into a grayscale image to reduce channel redundancy and highlight structural texture. Then, a noise filtering strategy based on the median filtering algorithm is used to smooth isolated pixels in the image caused by factors such as noise from the acquisition device and reflections from the steel cage. Subsequently, an edge sharpening algorithm based on the Laplacian operator is called to perform gradient enhancement on the image boundary region to improve the intensity contrast between the character edges and the background. For example, when processing a QR code image with a strong reflective area, the above steps can effectively highlight the boundaries of the QR code module and suppress stray reflections, ultimately obtaining an enhanced image with good contour clarity, which serves as the base image source for subsequent structural standardization processing.

[0050] S32: Perform size normalization, angle correction, and distortion correction operations on the enhanced identification code image to ensure that the image has a uniform recognition structure under different acquisition conditions, and finally obtain the preprocessed identification code image.

[0051] Specifically, based on the image boundary recognition results, the size normalization algorithm is first called to scale the identification code image proportionally to a uniform recognition resolution size to ensure the consistency of the model input scale. Then, the edge lines of the identification code are extracted based on the Hough transform to calculate the tilt angle, and the angle is corrected through rotation transformation. Subsequently, geometric correction is performed on the perspective distortion caused by the shooting angle. Four-point perspective mapping is used to calibrate the image edge points to a standard rectangular coordinate system. For example, in actual acquisition, a QR code image with a 15-degree tilt angle, slightly smaller size, and distortion in the lower right corner is obtained. After processing, it is unified to 300×300 pixels, the tilt angle is adjusted to 0 degrees, and the distortion is corrected to a rectangular area, which effectively improves the structural robustness and recognition accuracy of the subsequent recognition model.

[0052] In one embodiment, such as Figure 5 As shown, this method for tracking steel cage identification codes also includes: S401: Collect a dataset of identification code image samples and manually annotate the dataset. The annotation labels are used to indicate whether the identification code image can be successfully recognized during the actual decoding process. The dataset of identification code image samples includes recognizable image samples and unrecognizable image samples.

[0053] Specifically, by screening and organizing the QR code images retained from historical printing tasks, an image sample dataset covering various acquisition conditions and printing states is constructed. Then, human evaluators judge and label each image to determine whether it can be correctly recognized by the decoding software. The label is set as "1" to indicate successful recognition and "0" to indicate recognition failure. The image samples must include cases where the characters are blurred due to low nozzle pressure, the QR code is obscured by rust on the steel surface, or the image is severely skewed and distorted, which are unrecognizable. They must also include recognizable image samples with normal printing, uniform ambient light, and clear acquisition. For example, a sample set contains 5,000 images from 20 construction sites. About 15% of the images are labeled as unrecognizable samples due to blurry printing under high humidity conditions, while the remaining images are labeled as recognizable image samples.

[0054] S402: An image recognition rate evaluation model is constructed using a convolutional neural network structure, and the image recognition rate evaluation model is trained in a supervised manner using labeled signals to obtain a pre-trained image recognition rate evaluation model.

[0055] Specifically, following the standard convolutional neural network model construction process, an image recognition rate evaluation network structure was designed, comprising convolutional layers, pooling layers, batch normalization layers, and fully connected layers. The input is a normalized identification code image, and the output is a recognition probability score. During the model training phase, labeled images are used as the supervision target, and iterative optimization is performed using the cross-entropy loss function. The Adam optimizer is used to update the model weights. For example, when training on 4200 recognizable images and 800 unrecognizable images in the above sample dataset, the final converged model can achieve a classification performance of over 87% accuracy and an AUC value of 0.91 on the validation set, demonstrating the ability to quickly pre-evaluate the quality of images collected in the future.

[0056] In one embodiment, such as Figure 6 As shown, in step S50, before comparing the identification confidence score with the preset identification threshold, this method for tracking steel cage identification codes further includes: S501: Collects environmental state parameters of the identification code image and dynamically adjusts the preset recognition threshold based on the environmental state parameters. The environmental state parameters include the average image brightness, image clarity index, and occlusion degree index.

[0057] Specifically, the environmental state parameters of the identification code image are collected. The average image brightness is calculated by extracting the grayscale distribution of the current image frame to assess the lighting conditions of the scene. The image gradient is calculated based on the Laplacian operator, and its amplitude is aggregated to measure the overall image sharpness. Then, the background of the steel cage and the identification code area in the image are separated by semantic segmentation to count the proportion of their area obscured by foreign objects, dust, reflections, etc. Combining the above three indicators, an environmental state vector is constructed. This state vector is then compared with environmental state labels extracted from multiple sets of historical successful and failed recognition cases to dynamically adjust the recognition threshold. For example, when the image brightness is detected to be below 80 and the sharpness index is below 0.25 during the recognition process, the image recognition confidence threshold is lowered from 0.7 to 0.55 to improve the recognition error tolerance in the case of blurred edges. If the occlusion degree exceeds 35%, the backup matching strategy is triggered to continue trying, ensuring that there is still a certain recognition rate and judgment accuracy in cases such as insufficient lighting at the construction site or pollution on the surface of the rebar cage. For example, in a certain working condition, the image brightness is 63, the sharpness is 0.22, and the occlusion degree is 38%. The dynamically adjusted threshold ensures that the image can still be correctly recognized and a reliable identification code can be output.

[0058] In one embodiment, such as Figure 7 As shown, in step S70, based on the movement trajectory information, it is determined whether there is a path anomaly, which specifically includes: S71: Based on the target construction unit information corresponding to the identification code, a target path template is generated by matching from the preset path rules. The target path template is used to indicate the standard transfer path of the rebar cage.

[0059] Specifically, the target construction unit code embedded in the identification code is parsed and the corresponding construction unit attribute information is searched in the database, including the construction area number, floor number, and construction stage. Based on this attribute information, a matching path template is selected from the preset path rule set. The path template consists of multiple standard flow nodes, and the node order and time window are controlled by the construction plan and material scheduling constraints. For example, when the target construction unit indicated by the identification code is "D3-Tower B-10F", the matching path template includes the standard path "Processing Workshop A → Buffer Area B → Construction Tower Crane Discharge Port → Floor 10 Storage Point", and is accompanied by the expected arrival time range and transportation method label for subsequent path comparison analysis.

[0060] S72: Perform similarity matching analysis between the movement trajectory information and the target path template. If the matching similarity is lower than the anomaly judgment threshold, or if there are abnormal node events in the movement trajectory information, it is determined that the current rebar cage has a path anomaly. Abnormal node events include path interruption, frequent backtracking, or exceeding the construction area boundary.

[0061] Specifically, the historical positioning trajectory of the rebar cage throughout the entire process is analyzed, and its node sequence and dwell time information are extracted. The similarity is matched with the standard node sequence in the target path template using dynamic time warping. At the same time, the spatial location clustering of the offset points in the trajectory is used to determine whether there are abnormal behaviors such as boundary crossing or repeated entry. If the similarity value is lower than the set threshold of 0.65 or if the trajectory shows more than 3 backtracking behaviors or abrupt changes in path nodes, it is marked as a path anomaly. For example, if a rebar cage detours to the living area on its way to the construction area D3 and returns to the buffer area B twice to re-enter, the similarity drops to 0.52 and the anomaly recording mechanism is triggered for subsequent tracking and processing.

[0062] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0063] In one embodiment, a rebar cage identification code tracking system is provided, which corresponds one-to-one with the rebar cage identification code tracking method described in the above embodiments. For example... Figure 8 As shown, this rebar cage identification code tracking system includes an identification code generation module, an identification code printing and binding module, an image preprocessing module, an image evaluation module, a confidence level comparison module, a trajectory acquisition module, and a path anomaly detection module. Detailed descriptions of each functional module are as follows: The identification code generation module is used to obtain the steel cage production order information and generate the corresponding identification code based on the steel cage production order information; The identification code printing and binding module is used to perform identification code printing operation according to the identification code after the steel cage is manufactured. After printing, the identification code is bound to the preset RFID tag on the steel cage for a unique identifier for subsequent tracking and recording. The image preprocessing module is used to acquire the identification code image on the steel cage after the printing is completed, and to perform image preprocessing on the identification code image to obtain the preprocessed identification code image. The image evaluation module is used to evaluate the pre-processed identification code image using a pre-trained evaluation model to obtain a recognition confidence score. The confidence comparison module is used to compare the recognition confidence score with the preset recognition threshold. If the recognition confidence score is lower than the preset recognition threshold, the subsequent steel cage transfer steps are paused and the reprinting operation is performed until the recognition confidence score reaches the preset recognition threshold. The trajectory acquisition module is used to collect RFID tag data in real time during the transfer of the steel cage based on the RFID tag bound to the identification code, using a multi-source reader and writer device, and to construct the movement trajectory information of the steel cage based on the RFID tag data. The path anomaly detection module is used to determine whether there is a path anomaly based on the movement trajectory information. When an abnormal behavior is detected, an anomaly marker will be triggered, and anti-counterfeiting information will be bound to the tracking record associated with the identification code.

[0064] Optionally, the identifier generation module includes: The production information extraction submodule is used to extract the production information of the steel cage from the steel cage production order information. The production information includes component number, production batch, mold number and target construction unit information. The identification code generation submodule is used to combine production information according to preset encoding rules to generate a two-dimensional display code with embedded semantic fields, which serves as an identification code to uniquely identify the steel cage.

[0065] Optionally, the identification code printing binding module includes: The initial parameter generation submodule is used to generate initial printing parameters based on the identification code. The printing parameters include printing power, nozzle pressure and character spacing. The printing parameter adjustment submodule is used to collect surface environmental parameters of the rebar cage, dynamically adjust the initial printing parameters based on the surface environmental parameters to obtain the final printing parameters, and perform the identification code printing operation according to the final printing parameters. The surface environmental parameters include temperature and humidity, material type and surface roughness.

[0066] Optionally, the image preprocessing module includes: The image enhancement submodule is used to perform image grayscale conversion, noise filtering and edge sharpening operations on the identification code image to enhance the clarity and contour contrast of the identification code image, and obtain the enhanced identification code image. The image normalization submodule is used to perform size normalization, angle correction, and distortion correction operations on the enhanced identification code image to ensure that the image has a uniform recognition structure under different acquisition conditions, and finally obtain the preprocessed identification code image.

[0067] Optionally, the path anomaly detection module includes: The path template matching submodule is used to generate a target path template by matching the target construction unit information corresponding to the identifier code from the preset path rules. The target path template is used to indicate the standard circulation path of the rebar cage. The trajectory similarity analysis submodule is used to perform similarity matching analysis between the movement trajectory information and the target path template. If the matching similarity is lower than the anomaly judgment threshold, or if there are abnormal node events in the movement trajectory information, it is determined that the current rebar cage has a path anomaly. Abnormal node events include path interruption, frequent backtracking, or exceeding the construction area boundary.

[0068] Optionally, this steel cage identification code tracking system also includes: The image sample acquisition module is used to acquire the identification code image sample dataset and manually annotate the identification code image sample dataset. The annotation labels are used to indicate whether the identification code image can be successfully recognized in the actual decoding process. The identification code image sample dataset includes recognizable image samples and unrecognizable image samples. The evaluation model training module is used to construct an image recognition rate evaluation model using a convolutional neural network structure, and to perform supervised training on the image recognition rate evaluation model using labeled signals to obtain a pre-trained image recognition rate evaluation model.

[0069] The threshold adjustment module is used to collect the environmental state parameters of the identification code image and dynamically adjust the preset recognition threshold based on the environmental state parameters. The environmental state parameters include the average image brightness, image sharpness index, and occlusion degree index.

[0070] For specific limitations regarding a rebar cage identification code tracking system, please refer to the limitations of a rebar cage identification code tracking method described above, which will not be repeated here. Each module in the aforementioned rebar cage identification code tracking system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0071] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0072] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for tracking steel cage identification codes, characterized in that, The method for tracking steel cage identification codes includes: Obtain the steel cage production order information and generate a corresponding identification code based on the steel cage production order information; After the steel cage is manufactured, the identification code is printed according to the identification code. After the printing is completed, the identification code is bound to the RFID tag preset on the steel cage for use as a unique identifier for subsequent tracking and recording. After the printing of the steel cage is completed, the image of the identification code on the steel cage is captured, and the image of the identification code is preprocessed to obtain the preprocessed identification code image. The preprocessed identification code image is evaluated using a pre-trained evaluation model to obtain a recognition confidence score; The identification confidence score is compared with a preset identification threshold. If the identification confidence score is lower than the preset identification threshold, the subsequent steel cage transfer steps are paused and a reprinting operation is performed until the identification confidence score reaches the preset identification threshold. During the transfer of the steel cage, based on the RFID tag bound to the identification code, the data of the RFID tag is collected in real time using a multi-source reader / writer device, and the movement trajectory information of the steel cage is constructed based on the data of the RFID tag. Based on the movement trajectory information, it is determined whether there is any path abnormality. When abnormal behavior is determined, an abnormality marker will be triggered, and the anti-counterfeiting information will be bound to the tracking record associated with the identification code.

2. The method for tracking steel cage identification codes according to claim 1, characterized in that, The step of generating a corresponding identification code based on the steel cage production order information specifically includes: Extract the production information of the steel cage from the steel cage production order information. The production information includes component number, production batch, mold number, and target construction unit information. The production information is combined according to a preset encoding rule to generate a two-dimensional display code with embedded semantic fields, which serves as the identification code to uniquely identify the steel cage.

3. The method for tracking steel cage identification codes according to claim 1, characterized in that, The step of performing the identification code printing operation based on the identification code specifically includes: Based on the identification code, initial printing parameters are generated, including printing power, nozzle pressure and character spacing. The surface environment parameters of the reinforcing cage are collected, and the initial printing parameters are dynamically adjusted based on the surface environment parameters to obtain the final printing parameters. The identification code printing operation is then performed according to the final printing parameters. The surface environment parameters include temperature and humidity, material type, and surface roughness.

4. The method for tracking steel cage identification codes according to claim 1, characterized in that, The step of performing image preprocessing on the identification code image to obtain a preprocessed identification code image specifically includes: The identification code image is subjected to image grayscale conversion, noise filtering and edge sharpening operations to enhance the clarity and contour contrast of the identification code image, resulting in an enhanced identification code image. The enhanced identification code image is subjected to size normalization, angle correction, and distortion correction operations to ensure that the image has a uniform recognition structure under different acquisition conditions, and finally the preprocessed identification code image is obtained.

5. The method for tracking steel cage identification codes according to claim 1, characterized in that, The method for tracking steel cage identification codes also includes: Collect a dataset of identification code image samples and manually annotate the dataset. The annotation labels are used to indicate whether the identification code image can be successfully recognized during the actual decoding process. The dataset of identification code image samples includes recognizable image samples and unrecognizable image samples. An image recognition rate evaluation model is constructed using a convolutional neural network structure, and the image recognition rate evaluation model is trained in a supervised manner using the labeled tags as supervision signals to obtain the pre-trained image recognition rate evaluation model.

6. The method for tracking steel cage identification codes according to claim 1, characterized in that, Before comparing the identification confidence score with a preset identification threshold, the rebar cage identification code tracking method further includes: The environmental state parameters of the identification code image are collected, and the preset recognition threshold is dynamically adjusted based on the environmental state parameters. The environmental state parameters include the average image brightness, image clarity index, and occlusion degree index.

7. The method for tracking steel cage identification codes according to claim 1, characterized in that, The step of determining whether there is a path anomaly based on the movement trajectory information specifically includes: Based on the target construction unit information corresponding to the identification code, a target path template is generated by matching from the preset path rules. The target path template is used to indicate the standard transfer path of the steel cage. The movement trajectory information is matched with the target path template for similarity analysis. If the similarity is lower than the anomaly judgment threshold, or if there are abnormal node events in the movement trajectory information, it is determined that the current rebar cage has a path anomaly. The abnormal node events include path interruption, frequent backtracking, or exceeding the boundary of the construction area.

8. A steel cage identification code tracking system, characterized in that, The steel cage identification code tracking system includes: The identification code generation module is used to obtain steel cage production order information and generate a corresponding identification code based on the steel cage production order information. The identification code printing and binding module is used to perform identification code printing operation according to the identification code after the steel cage is manufactured, and bind the identification code to the preset RFID tag on the steel cage after printing, which is used as a unique identifier for subsequent tracking and recording. The image preprocessing module is used to acquire the identification code image on the steel cage after the steel cage is printed, and to perform image preprocessing on the identification code image to obtain the preprocessed identification code image. The image evaluation module is used to evaluate the preprocessed identification code image using a pre-trained evaluation model to obtain a recognition confidence score. The confidence comparison module is used to compare the recognition confidence score with a preset recognition threshold. If the recognition confidence score is lower than the preset recognition threshold, the subsequent steel cage transfer steps are paused and a reprinting operation is performed until the recognition confidence score reaches the preset recognition threshold. The trajectory acquisition module is used to collect data from RFID tags bound to the identification code in real time during the transfer of the steel cage using a multi-source reader / writer device, and to construct the movement trajectory information of the steel cage based on the data from the RFID tags. The path anomaly detection module is used to determine whether there is a path anomaly based on the movement trajectory information. When an abnormal behavior is detected, an anomaly marker will be triggered, and the anti-counterfeiting information will be bound to the tracking record associated with the identification code.

Images