System and method for determining position states of furnace door cross iron and clamping groove

Abstract

The invention provides a system and a method for determining the position state of a furnace door cross iron and a clamping groove, and the system comprises a control signal analysis module which is used for receiving a furnace door control signal sent by a PLC, and generating an image collection instruction based on the furnace door control signal; the image acquisition module responds to the image acquisition instruction and acquires a furnace door image; the position state detection module is used for inputting the furnace door image into an image detection model to obtain the position state and confidence of the furnace door cross iron and the clamping groove, and the position state is a furnace door lifting state or a furnace door closing state; the real-time pressure acquisition module responds to the detection completion signal and receives a real-time pressure value sent by the PLC; and the position state determination module is used for determining the position state as a target position state when the position state is a furnace door lifting state or a furnace door closing state, the confidence coefficient is greater than a first confidence coefficient threshold value and the real-time pressure value is greater than a pressure threshold value. The detection accuracy and efficiency of the position state of the transverse iron and the clamping groove are improved.

Description

Technical Field

[0001] This invention relates to the field of furnace door removal and closing control technology, specifically to a system and method for determining the position and state of the furnace door crossbar and the slot. Background Technology

[0002] The doors of the coke oven carbonization chamber need to be removed and closed during use. The specific process involves operators using a sliding door machine on the control panel of the coke pusher and coke quencher to lift the coke oven door upwards until the horizontal iron on the door is removed from its slot on the coke oven carbonization chamber door frame. Then, the door is opened and removed. Alternatively, the sliding door machine and the door can be slowly lowered together until the horizontal iron is embedded in the slot. During door removal, if the sliding door machine does not lift the door high enough, the horizontal iron may not be completely removed from its slot. Pulling the door open in this case can deform the door frame, ultimately preventing the door from closing properly or causing a serious safety hazard and environmental pollution incident. During door closing, if the horizontal iron does not fall into its slot, it indicates that the door is not closed tightly, posing a significant safety hazard as the door may collapse outwards.

[0003] Currently, operators of the coke pusher and coke quencher cars determine whether the furnace door can be removed and whether it is properly closed by visually observing whether the crossbar has moved out of or into its slot. However, in reality, the operators are located in the control room of the coke pusher and coke quencher car, at least 7 meters away from the furnace door. Furthermore, the crossbar and slot are relatively small structural components, and they are constantly exposed to high temperatures, oil, and dust. Tar and dust from the coking chamber accumulate on the crossbar and slot, further increasing the difficulty of visual inspection. Moreover, prolonged observation can lead to eye fatigue and misjudgments about whether the crossbar has moved out of or into its slot. This results in decreased efficiency in opening and closing the doors, disrupting production and posing safety hazards.

[0004] Therefore, there is an urgent need to provide a system and method for determining the position of the horizontal iron and the slot of the furnace door, so as to eliminate the technical problem that the accuracy and efficiency of the judgment result are low due to the subjective judgment of the operator on the position of the horizontal iron and the slot. Summary of the Invention

[0005] In view of this, it is necessary to provide a system and method for determining the position of the furnace door crossbar and the slot, so as to solve the technical problem that the existing technology relies solely on the operator's subjective judgment of the position of the crossbar and the slot, resulting in low accuracy and efficiency of the judgment results.

[0006] To address the aforementioned technical problems, in a first aspect, the present invention provides a method for determining the positional state of the furnace door crossbar and the slot, comprising: The control signal parsing module is used to receive the furnace door control signal sent by the PLC and generate an image acquisition command based on the furnace door control signal; The image acquisition module is used to acquire images of the furnace door in response to the image acquisition command; The position state detection module is used to input the furnace door image into the image detection model to obtain the position state and confidence level of the furnace door crossbar and the slot, wherein the position state is the furnace door lifted state or the furnace door closed state. The real-time pressure acquisition module is used to receive the real-time pressure value sent by the PLC in response to the detection completion signal. The position state determination module is used to determine the position state as the target position state when the position state is the furnace door raised or the furnace door closed, the confidence level is greater than a first confidence threshold, and the real-time pressure value is greater than a pressure threshold.

[0007] In one possible implementation, the furnace door control signal includes an operation command signal indicating whether the sliding door machine has completed a lifting or closing action, a hydraulic rod pressure value indicating when the sliding door machine performs a lifting or closing action, and a heartbeat packet indicating whether the communication link is normal; then the control signal parsing module includes a heartbeat packet detection unit, an action judgment unit, and a pressure judgment unit. The heartbeat packet detection unit is used to determine whether the communication link is normal based on the heartbeat packet. The action judgment unit is used to determine whether the sliding door machine has completed the picking action or closing action based on the furnace door control signal when the communication link is normal. The pressure judgment unit is used to determine whether the pressure value of the hydraulic rod is greater than the pressure threshold when the sliding door machine has completed the picking or closing action. If so, an image acquisition command is generated.

[0008] In one possible implementation, the image detection model includes a backbone subnetwork, a feature fusion subnetwork, and a single detection head. The backbone subnetwork is used to extract features from the furnace door image to obtain multi-scale features. The feature fusion subnetwork is used to fuse the multi-scale features to obtain fused features. The detection head is used to determine the position state and confidence level based on the fused features.

[0009] In one possible implementation, the position state detection module includes a preliminary detection unit and a non-maximum suppression unit; The preliminary detection unit is used to input the furnace door image into the image detection model to obtain multiple detection boxes and the initial confidence level of each detection box; The nonmaximum suppression unit is used to select the detection box with the highest initial confidence as the target detection box, calculate the cross-union ratio (CUP) of the target detection box and other detection boxes, remove detection boxes with CUP greater than the CUP threshold, select the detection result corresponding to the preferred detection box with the highest initial confidence among the remaining detection boxes as the position state, and use the initial confidence of the preferred detection box as the confidence.

[0010] In one possible implementation, when the image acquisition module includes multiple cameras, the number of furnace door images is multiple, and the position state includes multiple position sub-states corresponding one-to-one with the multiple cameras. In this case, the position state determination system of the furnace door crossbar and the slot further includes a multi-channel fusion processing module; the multi-channel fusion processing module includes a geometric constraint unit, a confidence constraint unit, and a filtering unit. The geometric constraint unit is used to calculate the center coordinates of the detection box corresponding to each of the position sub-states, and to determine whether the distance difference between the center coordinates and the theoretical center coordinates is greater than a preset distance. If it is greater, the detection box is removed to obtain at least one candidate detection box. The confidence constraint unit is used to determine whether the confidence of each candidate detection box is greater than the second confidence threshold. If it is greater, it is retained to obtain at least one detection box to be evaluated. The filtering unit is used to select the detection box with the highest confidence among the at least one detection boxes to be evaluated as the final selection detection box, and to use the detection result of the final selection detection box as the position state.

[0011] In one possible implementation, an image preprocessing module is also included, which is used to perform anomaly detection, size scaling, and normalization processing on the furnace door image.

[0012] In one possible implementation, a retry module is also included, which is used to control the image acquisition module and the position state detection module to work again when the detection result of the position state detection module is empty.

[0013] In one possible implementation, an output module is also included, which is used to send the target position status to the PLC so that the PLC can execute the next operation instruction based on the target position status.

[0014] In one possible implementation, the image detection model is trained based on a training set, which includes original furnace door image samples collected on-site from the coke oven and enhanced image samples generated by adding simulated smoke obscuring noise to the original furnace door image samples.

[0015] Secondly, the present invention also provides a method for determining the position state of the furnace door crossbar and the slot, including: Receive the furnace door control signal sent by the PLC, and generate an image acquisition command based on the furnace door control signal; In response to the image acquisition command, an image of the furnace door is acquired; The furnace door image is input into the image detection model to obtain the position status and confidence level of the furnace door crossbar and the slot. The position status is either the furnace door lifted state or the furnace door closed state. The system responds to the detection completion signal and receives the real-time pressure value sent by the PLC. When the position state is either furnace door raised or furnace door closed, the confidence level is greater than the first confidence threshold, and the real-time pressure value is greater than the pressure threshold, the position state is determined to be the target position state.

[0016] The beneficial effects of this invention are as follows: The furnace door crossbar and slot position determination system provided by this invention is a complete automated system consisting of a control signal analysis module, an image acquisition module, a position state detection module, and a position state determination module. It completely replaces the traditional method of judgment relying on operator visual observation. This fundamentally eliminates subjective misjudgments caused by factors such as distance, harsh environment, and eye fatigue, greatly improving the accuracy and efficiency of judgment. Furthermore, it can quickly and accurately determine whether the furnace door crossbar has completely moved out of or fallen into the slot, thus effectively preventing accidents caused by the crossbar not being completely disengaged before removing the furnace door and effectively preventing the risk of furnace door collapse caused by the crossbar not being in place after closing the furnace door. This directly eliminates major equipment damage and safety accidents caused by this, and avoids environmental problems such as harmful gas leaks caused by improperly closed furnace doors.

[0017] Furthermore, before the position status detection module, the furnace door control signal is first analyzed by the control signal analysis module. The furnace door image is then acquired and detected only when the image acquisition command is generated. Compared with the technical solution without a control signal analysis module, this achieves the determination of the image acquisition timing, avoids the problem of judging a large number of furnace door images and the inability to obtain a stable furnace door image during movement, which leads to poor accuracy of the detected position status. This further improves the detection efficiency and accuracy of the position status.

[0018] Furthermore, the position status determination module uses real-time pressure values ​​as an auxiliary criterion when determining the target position status. This avoids the problem that even if the mechanical action is completed, insufficient pressure may result in the horizontal bar not completely disengaging from or falling into the slot. In other words, pressure verification can eliminate false completion signals caused by equipment aging, hydraulic leakage, or other reasons, further ensuring the reliability of the determined target position status. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic diagram of an embodiment of the furnace door crossbar and slot position determination system provided by the present invention; Figure 2 This is a schematic diagram of an embodiment of the control signal parsing module provided by the present invention; Figure 3 This is a schematic diagram of an embodiment of the image detection model provided by the present invention; Figure 4 A schematic diagram of an embodiment of the position state detection module provided by the present invention; Figure 5 A schematic diagram of an embodiment of the multi-channel fusion processing module provided by the present invention; Figure 6 This is a schematic flowchart of an embodiment of the method for determining the position and state of the furnace door crossbar and the slot provided by the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this invention illustrate operations implemented according to some embodiments of the invention. It should be understood that the operations in the flowcharts may be implemented out of order, and steps without logical contextual relationships may be reversed or performed simultaneously. Furthermore, those skilled in the art, guided by the content of this invention, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0023] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0024] This invention provides a system and method for determining the position and state of the crossbar and the slot of the furnace door, which will be described below.

[0025] Figure 1 A schematic diagram of an embodiment of the furnace door crossbar and slot position determination system provided by the present invention is shown below. Figure 1 As shown, the furnace door crossbar and slot position determination system 10 includes: The control signal parsing module 100 is used to receive the furnace door control signal sent by the PLC and generate image acquisition instructions based on the furnace door control signal.

[0026] Among them, the furnace door control signal includes at least the operation command signal, which is generated when the sliding door machine has completed the lifting or closing action, to ensure that the sliding door machine completes the sliding action.

[0027] The image acquisition module 200 is used to acquire images of the furnace door in response to image acquisition commands.

[0028] Because coke oven production generates coke oven gas, which is a flammable and explosive gas, and the temperature and dust concentration in the oven door area are extremely high when the coke pusher is working, accurate image acquisition of the oven door is required under these conditions. In a specific embodiment of the present invention, the camera in the image acquisition module 200 is a stainless steel explosion-proof air-cooled integrated camera.

[0029] The position state detection module 300 is used to input the furnace door image into the image detection model to obtain the position state and confidence level of the furnace door crossbar and the slot. The position state is either the furnace door is in the lifted state or the furnace door is in the closed state.

[0030] Among them, the image detection model can be a general target detection model, such as the YOLO series models.

[0031] The real-time pressure acquisition module 400 is used to receive the real-time pressure value sent by the PLC in response to the detection completion signal.

[0032] Specifically, when the position status detection module 300 obtains the position status and confidence level, it immediately generates a detection completion signal.

[0033] The position status determination module 500 is used to determine the position status as the target position status when the position status is the furnace door raised or the furnace door closed, the confidence level is greater than the first confidence level threshold, and the real-time pressure value is greater than the pressure threshold.

[0034] It should be understood that both the first confidence threshold and the stress threshold can be set or adjusted according to the application scenario, and no specific limitations are made here.

[0035] Compared with existing technologies, the furnace door crossbar and slot position determination system 10 provided in this embodiment of the invention is a complete automated system composed of a control signal analysis module 100, an image acquisition module 200, a position status detection module 300, and a position status determination module 500. It completely replaces the traditional method of judgment relying on operator visual observation. This fundamentally eliminates subjective misjudgments caused by factors such as distance, harsh environment, and eye fatigue, greatly improving the accuracy and efficiency of judgment. Furthermore, it can quickly and accurately determine whether the furnace door crossbar has completely moved out of or fallen into the slot, thus effectively preventing accidents caused by the crossbar not being completely disengaged before removing the furnace door, and effectively preventing the risk of furnace door collapse caused by the crossbar not being in place after closing the furnace door. This directly eliminates major equipment damage and safety accidents caused by this, and avoids environmental problems such as harmful gas leaks caused by improperly closed furnace doors.

[0036] Furthermore, before the position status detection module 300, the furnace door control signal is first analyzed by the control signal analysis module 100. The furnace door image is then acquired and detected only when the image acquisition command is generated. Compared with the technical solution without setting a control signal analysis module, this achieves the determination of the image acquisition timing, avoids the problem of judging a large number of furnace door images and the inability to obtain a stable furnace door image during movement, which leads to poor accuracy of the determined position status. This further improves the detection efficiency and accuracy of the position status.

[0037] Furthermore, the position status determination module 500 uses real-time pressure values ​​as an auxiliary judgment criterion when determining the target position status. This avoids the problem that even if the mechanical action is completed, insufficient pressure may result in the horizontal bar not completely disengaging from or falling into the slot. In other words, pressure verification can eliminate false completion signals caused by equipment aging, hydraulic leakage, etc., further ensuring the reliability of the determined target position status.

[0038] When a sliding door operator opens or closes the furnace door, it consists of mechanical and pressure actions. The mechanical action is responsible for moving the horizontal bar out of or into the slot, while the pressure action ensures that it can be completely moved out or inserted under the specified mechanical action. For example, if the mechanical action is completed, but the hydraulic pressure of the hydraulic action does not reach the theoretical pressure, the horizontal bar cannot be completely disengaged from or inserted into the slot.

[0039] Therefore, to ensure the accuracy of image acquisition, in some embodiments of the present invention, the furnace door control signal includes an operation command signal indicating whether the sliding door machine has completed the lifting or closing action, a hydraulic rod pressure value indicating when the sliding door machine performs the lifting or closing action, and a heartbeat packet indicating whether the communication link is normal; then, as Figure 2 As shown, the control signal analysis module 100 includes a heart rate detection unit 110, an action judgment unit 120, and a pressure judgment unit 130; The heartbeat detection unit 110 is used to determine whether the communication link is normal based on the heartbeat packet. The action judgment unit 120 is used to determine whether the sliding door machine has completed the picking action or closing action based on the furnace door control signal when the communication link is normal.

[0040] Specifically, when the sliding door machine completes the picking action, it can generate a furnace door control signal of "picking completed"; when the sliding door machine completes the closing action, it can generate a furnace door control signal of "closing completed". Based on these two types of signals, it is possible to intuitively and quickly determine whether the sliding door machine has completed the picking or closing action.

[0041] The pressure judgment unit 130 is used to determine whether the hydraulic rod pressure value is greater than the pressure threshold when the sliding door motor has completed the picking or closing action. If so, an image acquisition command is generated.

[0042] This embodiment of the invention, by setting up a heartbeat detection unit 110, ensures that the PLC status upon which the system makes decisions is up-to-date and reliable, avoiding false triggering due to communication interruptions. The pressure verification by the pressure judgment unit 130 compensates for the inadequacy of the purely mechanical position signal in the action judgment unit 120, preventing forced operation under insufficient force and protecting the furnace door. In other words, by combining communication status, mechanical position, and hydraulic pressure signals, misjudgments caused by errors in a single signal are avoided, thereby improving the accuracy of image acquisition command generation timing.

[0043] The YOLO series of target detection models employs a multi-scale detection head architecture, comprising three detection heads of different scales, responsible for detecting targets of large, medium, and small sizes, respectively. While this design can adapt to the detection needs of targets of different sizes, in the application scenario of furnace door crossbars and slots, the dimensions of the crossbars and slots are relatively fixed. If the detection head architecture is still used, computational redundancy and high model complexity will result in unnecessary computational overhead, leading to low efficiency in determining the position state.

[0044] Therefore, in some embodiments of the present invention, such as Figure 3As shown, the image detection model includes a backbone sub-network, a feature fusion sub-network, and a single detection head. The backbone sub-network is used to extract features from the furnace door image to obtain multi-scale features. The feature fusion sub-network is used to fuse the multi-scale features to obtain fused features. The detection head is used to determine the position state and confidence level based on the fused features.

[0045] This invention embodiment uses only a single detection head, resulting in a simple model structure. Furthermore, since only a single detection head is used, the feature fusion path from the network is simplified, requiring only the feature extraction and fusion paths related to the single detection head to be retained.

[0046] It should be noted that this embodiment of the invention is an improvement on the structure of YOLOv8. YOLOv8 includes three detection heads corresponding to 8x, 16x, and 32x downsampling feature maps. This embodiment of the invention retains the detection head corresponding to the 16x downsampling feature map. The receptive field of the 16x downsampling feature map matches the detection target in this embodiment of the invention. Furthermore, at this scale, an optimal balance is achieved between detail information and semantic information.

[0047] Since the YOLOv8 architecture is relatively mature, the unmodified architecture will not be elaborated upon here; for details, please refer to [link to relevant documentation]. Figure 3 .

[0048] In the actual scenario of furnace door crossbars and slots, there is only one target in each image. Theoretically, the image detection model only needs to output one detection box. However, in practical applications, the image detection model outputs multiple detection boxes, which may overlap. Therefore, it is necessary to determine the final required detection box from among these multiple boxes to obtain the position state. In some embodiments of the present invention, such as... Figure 4 As shown, the position state detection module 300 includes a preliminary detection unit 310 and a non-maximum suppression unit 320; The preliminary detection unit 310 is used to input the furnace door image into the image detection model to obtain multiple detection boxes and the initial confidence of each detection box; The nonmaximum suppression unit 320 is used to take the detection box with the highest initial confidence as the target detection box, calculate the cross-union ratio (CUP) of the target detection box and other detection boxes, remove detection boxes with CUP greater than the CUP threshold, take the detection result corresponding to the preferred detection box with the highest initial confidence in the remaining detection boxes as the position state, and take the initial confidence of the preferred detection box as the confidence.

[0049] The embodiments of the present invention improve the accuracy and reliability of detection results by setting a nonmaximum suppression unit 320 to remove duplicate detection frames.

[0050] To avoid the technical problem of inaccurate results due to perspective or other factors in single-camera acquisition, in some embodiments of the present invention, when the image acquisition module 200 includes multiple cameras, the number of furnace door images is multiple, and the position state includes multiple position sub-states corresponding one-to-one with the multiple cameras. To determine the final accurate position state among the multiple position sub-states, in some embodiments of the present invention, such as... Figure 1 and Figure 5 As shown, the furnace door crossbar and slot position determination system 10 also includes a multi-channel fusion processing module 600; the multi-channel fusion processing module 600 includes a geometric constraint unit 610, a confidence constraint unit 620 and a filtering unit 630; The geometric constraint unit 610 is used to calculate the center coordinates of the detection box corresponding to each sub-state and to determine whether the distance difference between the center coordinates and the theoretical center coordinates is greater than a preset distance. If it is greater, the detection box is removed to obtain at least one candidate detection box. The confidence constraint unit 620 is used to determine whether the confidence of each candidate detection box is greater than the second confidence threshold. If it is greater, it is retained, and at least one detection box to be evaluated is obtained. The filtering unit 630 is used to select the detection box with the highest confidence among at least one detection box to be evaluated as the final selected detection box, and to use the detection result of the final selected detection box as the position state.

[0051] The selection criterion for geometric constraint unit 610 is as follows: Since the camera is fixed on the coking carriage, the relative position and angle between the camera and the furnace door crossbar / slot are basically fixed during each operation. Therefore, the area where the crossbar and slot appear in the image can be predicted. That is, the center coordinates should be within a certain theoretical range. Based on this, the center point coordinates of each detection box are calculated. If this coordinate deviates from the expected reasonable area, the detection result is considered to be a localization error or a false detection, and it is discarded.

[0052] It should be understood that if multiple cameras are configured, the decision logic of the position status determination module 500 can be expressed as follows:

[0053] In the formula, This indicates that the detection results from n cameras under the current furnace door are subjected to a logical AND operation. This represents the logical AND operation; P , P normal These represent the pressure provided by the hydraulic rod of the sliding door operator when operating the current furnace door, and the theoretical pressure that the hydraulic rod of the sliding door operator needs to provide when lifting or closing the furnace door. Let be the confidence level of the detection output of the i-th camera; This is the second confidence threshold.

[0054] It should be understood that the second confidence threshold is less than the first confidence threshold.

[0055] To improve the detection performance of the image detection model, in some embodiments of the present invention, such as... Figure 1 As shown, the furnace door horizontal iron and the card slot position status determination system 10 also includes an image preprocessing module 700, which is used to perform abnormality, size scaling and normalization processing on the furnace door image.

[0056] Specifically, the abnormality judgment can be made by judging whether indicators such as clarity meet the conditions. If they do not meet the conditions, the furnace door image is abnormal; otherwise, the furnace door image is normal.

[0057] Specifically, scaling means scaling the furnace door image to a size that matches the set size.

[0058] Specifically, the normalization process involves converting the pixel values ​​of the furnace door image from unsigned data to floating-point data, calculating the mean for each of the three channels of the image, and performing a subtraction operation on the corresponding mean for each channel. The calculation formula is as follows:

[0059] In the formula, These represent the height and width of the image after scaling, respectively; This indicates the image before normalization. c The first channel (one of the three RGB channels) i The pixel value of each pixel; This represents the conversion factor for converting unsigned data to floating-point data; This indicates that after image normalization, c The first channel (one of the three RGB channels) i The pixel value of each pixel; Because furnace door detection can be affected by smoke and other factors from the coke oven, resulting in ineffective detection of position status, and single image recognition attempts may fail due to various transient interferences, this invention aims to address environmental interference and improve the overall success rate and availability of the system. In some embodiments of this invention, such as... Figure 1 As shown, the furnace door crossbar and the card slot position status determination system 10 also includes a retry module 800. The retry module 800 is used to control the image acquisition module 200 and the position status detection module 300 to work again when the detection result of the position status detection module 300 is empty.

[0060] By setting up a retry module 800, this embodiment of the invention can provide a self-recovery opportunity for the detection process, thereby improving its detection success rate and availability.

[0061] To avoid dead loops caused by permanent malfunctions such as camera damage, a maximum number of retries should be set. When the maximum number of retries has not been reached, the image acquisition module 200 and the position status detection module 300 should be controlled to re-detect. When the maximum number of retries has been reached, an error should be reported.

[0062] In the entire coke pusher car control system, the PLC is responsible for coordinating all actuators such as the sliding door machine, coke pusher, and coal leveling rod, and handling communication with the upper-level MES system. Therefore, in some embodiments of the present invention, such as Figure 1 As shown, the furnace door crossbar and slot position determination system 10 also includes an output module 900, which is used to send the target position status to the PLC so that the PLC can execute the next operation instruction based on the target position status.

[0063] In this embodiment of the invention, the target position status is fed back to the PLC, which can then make subsequent decisions.

[0064] It should be understood that the image detection model needs to be trained before use to address the problem of smoke interference in this scenario, thereby further improving detection accuracy and robustness. In some embodiments of the present invention, the image detection model is trained based on a training set, which includes original furnace door image samples collected on-site from the coke oven and enhanced image samples generated by adding simulated smoke occlusion noise to the original furnace door image samples.

[0065] This invention improves the detection success rate under smoky conditions by adding simulated smoke occlusion noise to the original furnace door image samples during training.

[0066] In addition to the above-mentioned enhancement methods, in some embodiments of the present invention, the original furnace door image samples can also be enhanced by improving image brightness, contrast, positional offset, random enhancement, etc., in order to enhance the generalization ability of the model.

[0067] In summary, the method for determining the position and status of the furnace door crossbar and the slot proposed in this embodiment of the invention uses an image detection model to distinguish the state of the furnace door crossbar and the slot. Then, through system modules such as the control signal analysis module and the position and status detection module, the final result is fed back to the PLC, thereby determining whether the furnace door is properly lifted and closed during the lifting and closing process. This system can be directly integrated into the operation process of the coke pusher and coke quencher, improving operational efficiency and preventing safety accidents.

[0068] On the other hand, embodiments of the present invention also provide a method for determining the position and state of the furnace door crossbar and the slot, such as... Figure 6 As shown, the method for determining the position of the furnace door crossbar and the slot includes: S601: Receives the furnace door control signal sent by the PLC and generates an image acquisition command based on the furnace door control signal; S602: Respond to the image acquisition command and acquire the furnace door image; S603. Input the furnace door image into the image detection model to obtain the position status and confidence level of the furnace door horizontal iron and the slot. The position status is either the furnace door lifted state or the furnace door closed state. S604, respond to the detection completion signal and receive the real-time pressure value sent by the PLC; S605. When the position status is furnace door raised or furnace door closed, the confidence level is greater than the first confidence level threshold, and the real-time pressure value is greater than the pressure threshold, the position status is determined to be the target position status.

[0069] It should be noted that the method for determining the position of the furnace door horizontal iron and the slot provided in the above embodiments can realize the technical solution described in the above embodiments of the furnace door horizontal iron and slot position determination system. The principle or specific implementation details of each of the above steps can be found in the corresponding content in the above embodiments of the furnace door horizontal iron and slot position determination system, which will not be elaborated here.

[0070] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0071] The above provides a detailed description of the system and method for determining the position of the furnace door crossbar and the slot provided by the present invention. Specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A system for determining the position and state of a furnace door crossbar and a slot, characterized in that, include: The control signal parsing module is used to receive the furnace door control signal sent by the PLC and generate an image acquisition command based on the furnace door control signal; The image acquisition module is used to acquire images of the furnace door in response to the image acquisition command; The position state detection module is used to input the furnace door image into the image detection model to obtain the position state and confidence level of the furnace door crossbar and the slot, wherein the position state is the furnace door lifted state or the furnace door closed state. The real-time pressure acquisition module is used to receive the real-time pressure value sent by the PLC in response to the detection completion signal. The position state determination module is used to determine the position state as the target position state when the position state is the furnace door raised or the furnace door closed, the confidence level is greater than a first confidence threshold, and the real-time pressure value is greater than a pressure threshold.

2. The furnace door crossbar and slot position determination system according to claim 1, characterized in that, The furnace door control signal includes an operation command signal indicating whether the sliding door machine has completed the lifting or closing action, a hydraulic rod pressure value indicating when the sliding door machine performs the lifting or closing action, and a heartbeat packet indicating whether the communication link is normal; therefore, the control signal parsing module includes a heartbeat packet detection unit, an action judgment unit, and a pressure judgment unit. The heartbeat packet detection unit is used to determine whether the communication link is normal based on the heartbeat packet. The action judgment unit is used to determine whether the sliding door machine has completed the picking action or closing action based on the furnace door control signal when the communication link is normal. The pressure judgment unit is used to determine whether the pressure value of the hydraulic rod is greater than the pressure threshold when the sliding door machine has completed the picking or closing action. If so, an image acquisition command is generated.

3. The system for determining the position and state of the furnace door crossbar and the slot according to claim 1, characterized in that, The image detection model includes a backbone subnetwork, a feature fusion subnetwork, and a single detection head. The backbone subnetwork is used to extract features from the furnace door image to obtain multi-scale features. The feature fusion subnetwork is used to fuse the multi-scale features to obtain fused features. The detection head is used to determine the position state and confidence level based on the fused features.

4. The furnace door crossbar and slot position determination system according to claim 1, characterized in that, The position state detection module includes a preliminary detection unit and a non-maximum suppression unit; The preliminary detection unit is used to input the furnace door image into the image detection model to obtain multiple detection boxes and the initial confidence level of each detection box; The nonmaximum suppression unit is used to select the detection box with the highest initial confidence as the target detection box, calculate the cross-union ratio (CUP) of the target detection box and other detection boxes, remove detection boxes with CUP greater than the CUP threshold, select the detection result corresponding to the preferred detection box with the highest initial confidence among the remaining detection boxes as the position state, and use the initial confidence of the preferred detection box as the confidence.

5. The furnace door crossbar and slot position determination system according to claim 1, characterized in that, When the image acquisition module includes multiple cameras, the number of furnace door images is multiple, and the position status includes multiple position sub-states corresponding one-to-one with the multiple cameras. Then, the position status determination system of the furnace door crossbar and the slot also includes a multi-channel fusion processing module. The multi-path fusion processing module includes a geometric constraint unit, a confidence constraint unit, and a filtering unit; The geometric constraint unit is used to calculate the center coordinates of the detection box corresponding to each of the position sub-states, and to determine whether the distance difference between the center coordinates and the theoretical center coordinates is greater than a preset distance. If it is greater, the detection box is removed to obtain at least one candidate detection box. The confidence constraint unit is used to determine whether the confidence of each candidate detection box is greater than the second confidence threshold. If it is greater, it is retained to obtain at least one detection box to be evaluated. The filtering unit is used to select the detection box with the highest confidence among the at least one detection boxes to be evaluated as the final selection detection box, and to use the detection result of the final selection detection box as the position state.

6. The system for determining the position and state of the furnace door crossbar and the slot according to claim 1, characterized in that, The system also includes an image preprocessing module, which is used to perform anomaly detection, size scaling, and normalization on the furnace door image.

7. The furnace door crossbar and slot position determination system according to any one of claims 1-6, characterized in that, The system also includes a retry module, which is used to control the image acquisition module and the position state detection module to work again when the detection result of the position state detection module is empty.

8. The system for determining the position and state of the furnace door crossbar and the slot according to claim 1, characterized in that, The system also includes an output module, which sends the target position status to the PLC so that the PLC can execute the next operation instruction based on the target position status.

9. The system for determining the position of the furnace door crossbar and the slot according to claim 1, characterized in that, The image detection model is trained based on a training set, which includes original furnace door image samples collected on-site from the coke oven and enhanced image samples generated by adding simulated smoke obscuring noise to the original furnace door image samples.

10. A method for determining the positional state of a furnace door crossbar and a slot, characterized in that, include: Receive the furnace door control signal sent by the PLC, and generate an image acquisition command based on the furnace door control signal; In response to the image acquisition command, an image of the furnace door is acquired; The furnace door image is input into the image detection model to obtain the position status and confidence level of the furnace door crossbar and the slot. The position status is either the furnace door lifted state or the furnace door closed state. The system responds to the detection completion signal and receives the real-time pressure value sent by the PLC. When the position state is either furnace door raised or furnace door closed, the confidence level is greater than the first confidence threshold, and the real-time pressure value is greater than the pressure threshold, the position state is determined to be the target position state.

Images