Filter press filtrate water flow rate detection method, device, electronic equipment and storage medium

Abstract

The application provides a filter press filtrate water flow detection method and device, electronic equipment and storage medium, wherein: a drainage pipe detection model is used to process an RGB image of a filtrate monitoring area to obtain a detection result of each drainage pipe, thereby calculating the suction potential source strength of the outlet of each drainage pipe; a water flow segmentation network is used to process the RGB image to obtain a segmentation result of the filtrate water flow, thereby calculating the potential point feature of each filtrate water flow; the correlation potential field value of the drainage pipe and the water flow is calculated based on the suction potential source strength of the outlet of each drainage pipe and the potential point feature of each filtrate water flow, so as to determine the matched drainage pipe of each filtrate water flow; the input feature of each filtrate water flow is determined based on the segmentation result of each filtrate water flow, the detection result of the matched drainage pipe and the correlation potential field value; and a water flow estimation model is used to process the input feature of each filtrate water flow to obtain the flow of each filtrate water flow. The application can stably and reliably monitor the filtrate water flow with high precision.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method, apparatus, electronic device and storage medium for detecting the flow rate of filtrate in a filter press. Background Technology

[0002] Traditional machine vision monitoring solutions often employ fixed industrial cameras to capture images of the filter press drain pipe area. They then use conventional image processing algorithms such as fixed threshold segmentation, Canny edge detection, and contour matching to locate the drain pipe and extract the water flow area. The flow rate is then estimated by simply calculating the pixel area of ​​the water flow. This approach lacks specific optimization for the complex operating conditions of filter presses and suffers from several insurmountable technical shortcomings in actual operation. When faced with the tilting, interlacing, and overlapping pose changes of the drain pipes due to long-term equipment vibration and material impact, traditional algorithms lack spatial topological constraints and rely solely on external contour features. This fails to accurately distinguish between different drain pipes, frequently resulting in missed detections, false detections, incorrect numbering, and blurred boundaries, thus failing to provide a stable foundation for subsequent water flow matching.

[0003] For the identification of filtrate flow scenarios, filter press filtrate often exhibits characteristics such as being small, discontinuous, transparent, and having a color that converges with the equipment background. Furthermore, the outlets of densely packed drainage pipes are prone to merging, crossing, and splashing of water flows. Traditional segmentation algorithms rely on fixed grayscale thresholds and contour features, failing to distinguish transparent water flows from the background, and even unable to split merged water flows into instances. This results in incomplete water flow segmentation, excessive pseudo-contours, ineffective feature extraction, and even large-area missed segmentation. In the matching stage between water flow and drainage pipes, existing technologies only use spatial distance and vertical positional relationships for simple association, without introducing physical filtrate discharge logic and spatial constraints. In scenarios with intersecting pipes and merging water flows, mismatches, one-to-many, and many-to-one matching errors are highly likely. Simultaneously, they cannot effectively eliminate interference from invalid water flows such as splashes, residual water stains, and equipment dripping, making the reliability of the matching results completely unacceptable for industrial environments.

[0004] In the process of flow quantification and water flow color determination, existing technologies only estimate flow by statistically analyzing the pixel area of ​​the water flow, without combining multi-dimensional features such as the size of the drain pipe, the shape of the water flow, and the relative position of the pipe. As a result, the estimated flow values ​​have a large deviation and are not of practical reference value. Water flow color determination is based on simple RGB color threshold comparison, which is easily affected by factors such as changes in ambient light, equipment stains, and water reflection. It is impossible to accurately distinguish key states such as normal filtrate, filtrate containing impurities, and filtrate containing oil. Summary of the Invention

[0005] In view of this, this application provides a method, apparatus, electronic device and storage medium for detecting the flow rate of filtrate in a filter press, in order to solve the above-mentioned technical problems.

[0006] In a first aspect, this application proposes a method for detecting the flow rate of filtrate in a filter press. The filter press includes a drain pipe, with each drain pipe corresponding to a flow of filtrate, comprising: Acquire RGB images of the filtrate monitoring area of ​​the filter press under test; The RGB image is processed using a drainage pipe detection model to obtain the detection results of each drainage pipe; the drainage pipe outlet is set as the attraction potential source, and the attraction potential source intensity of each drainage pipe outlet is calculated based on the detection results. The RGB image is processed using a water flow segmentation network to obtain the water flow morphology features and segmentation results of each filtrate flow; the centroid of each filtrate flow is set as the attracted potential point, and the attracted potential point features of each filtrate flow are calculated based on the segmentation results. Based on the strength of the attraction potential source and the characteristics of the attraction potential point of the filtrate flow, the associated potential field value of each drain pipe and the filtrate flow is calculated; based on the associated potential field value, the drain pipe matched for each filtrate flow is determined. The input characteristics of each filtrate flow are determined based on the segmentation results, flow morphology characteristics, correlation potential field values, and detection results. Input the input features into the water flow estimation model to obtain the flow rate of the filtrate from the filter press to be tested.

[0007] Furthermore, based on the detection results, the intensity of the attraction potential source at each drain pipe outlet is calculated, including: The two-dimensional coordinates of each drainage pipe are determined based on the drainage pipe detection frame in the detection results; The intensity of the attraction potential source at each drainage pipe outlet is calculated based on two-dimensional coordinates.

[0008] Furthermore, based on the segmentation results, the attraction potential point characteristics of each filtrate flow are calculated, including: The pixel area of ​​each filtrate flow is determined based on the binarized segmentation mask in each segmentation result; The attraction potential point characteristics of each filtrate flow are calculated based on the pixel area.

[0009] Furthermore, the RGB images are processed using a flow segmentation network to obtain the flow morphology characteristics of each filtrate flow, including: The RGB image is processed using a water flow segmentation network to obtain the actual physical area of ​​each filtrate water flow, the aspect ratio of the water flow profile, and the perimeter of the water flow segmentation profile. The flow morphology characteristics of each filtrate flow are constructed by using the actual physical area, the aspect ratio of the flow profile, and the perimeter of the flow segmentation profile.

[0010] Furthermore, based on the correlated potential field value, the drain pipes matching each filtrate flow are determined, including: Corresponding to a flow of filtrate, match the associated potential field value of each drain pipe; The drain pipe corresponding to the maximum value in the associated potential field is determined as the drain pipe for the filtrate flow.

[0011] Furthermore, the methods also include: The successfully matched drain pipe is marked as being in the liquid discharge state, the drain pipe that is not matched with the filtrate flow is marked as being in the liquid-free state, and the flow without a corresponding matching drain pipe is automatically determined as an invalid flow.

[0012] Furthermore, based on the segmentation results, flow morphology characteristics, correlation potential field values, and detection results, the input characteristics of each filtrate flow are determined, including: The input features of each filtrate flow are obtained by dimensionally concatenating the water flow morphology features, the drainage pipe detection box in the segmentation results, and the associated potential field value.

[0013] Secondly, this application proposes a filter press filtrate flow rate detection device, comprising: The acquisition unit is used to acquire RGB images of the filtrate monitoring area of ​​the filter press to be tested; The first processing unit is used to process the RGB image using the drainage pipe detection model to obtain the detection results of each drainage pipe; the drainage pipe outlet is set as the attraction potential source, and the attraction potential source intensity of each drainage pipe outlet is calculated based on the detection results. The second processing unit is used to process the RGB image using a water flow segmentation network to obtain the water flow morphology features and segmentation results of each filtrate water flow; the centroid of each filtrate water flow is set as the attracted potential point, and the attracted potential point features of each filtrate water flow are calculated based on the segmentation results. The matching unit is used to calculate the associated potential field value between each drain pipe and the filtrate flow based on the strength of the attraction potential source and the characteristics of the attraction potential point of the filtrate flow; and to determine the drain pipe to be matched with each filtrate flow based on the associated potential field value. The determination unit is used to determine the input characteristics of each filtrate flow based on the segmentation results, water flow morphology characteristics, correlation potential field values, and detection results; The detection unit is used to input the input features into the water flow estimation model to obtain the flow rate of the filtrate from the filter press to be tested.

[0014] Thirdly, this application proposes an electronic device, including: a processor; and a memory storing a program, wherein the program includes instructions that, when executed by the processor, cause the processor to perform the method described above.

[0015] Fourthly, this application proposes a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the methods described above.

[0016] The beneficial effects of this application are as follows: This application designs a method for accurate identification of interlaced drainage pipes. By providing strong spatial constraints through the spatial topological feature field of drainage pipes, it can accurately distinguish drainage pipes in inclined, interlaced, and overlapping states. The false detection rate and false detection rate are significantly reduced, the detection results are stable and unique, and it has real-time pose adaptation capability. Attached Figure Description

[0017] Figure 1 A flowchart of a filter press filtrate flow rate detection method provided in an embodiment of this application; Figure 2 This is a schematic diagram of a drainage pipe inspection provided in an embodiment of this application; Figure 3 This is a schematic diagram of the spatial topological feature field of the drainage pipe provided in an embodiment of this application; Figure 4 This is a schematic diagram of the potential field associated with the water flow and drainage pipe provided in an embodiment of this application. Figure 5 This is a functional structure diagram of the filter press filtrate flow rate detection device provided in the embodiments of this application; Figure 6 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] First, a brief introduction to the design concept of the embodiments of this application will be given.

[0021] Filter presses are core industrial equipment in solid-liquid separation processes in coal, chemical, metallurgical, and environmental protection industries. The real-time discharge status, flow rate, and color of the filtrate directly reflect the sealing effect of the filter plates, the permeability of the filter cloth, and the quality of material separation. These are crucial for achieving fully automated feeding, plate closing, and cake discharge closed-loop control of the filter press. Currently, monitoring of filter press filtrate flow rate and discharge status in industrial settings mainly relies on two methods: manual inspection and traditional machine vision monitoring. Neither of these techniques is suitable for complex industrial conditions and exhibits significant technical deficiencies and application limitations.

[0022] Manual inspection is currently the most common monitoring method. Maintenance personnel periodically visit the filter press site to visually observe the liquid discharge status of multiple drain pipes, subjectively estimating the water flow rate and color, and manually recording the working status of each drain pipe. This method relies entirely on subjective human judgment, resulting in extremely low monitoring efficiency and an inability to achieve 24 / 7 uninterrupted real-time monitoring. It is also highly susceptible to misjudgments and omissions due to factors such as personnel fatigue, viewing angle deviations, changes in lighting, and nighttime operations. Furthermore, filter press drain pipes are typically densely arranged and staggered, making it impossible for manual personnel to quickly and accurately match the water flow with the corresponding drain pipe. Simultaneous and quantifiable data collection and recording of liquid discharge from multiple drain pipes is also impossible. The monitoring data lacks continuity and traceability, failing to provide reliable data support for the automated intelligent control of the filter press.

[0023] Existing monitoring technologies cannot solve the core problems of difficulty in distinguishing inclined and intersecting drainage pipes, difficulty in separating irregular / merging water flows, difficulty in matching water flow with drainage pipes, and inaccurate judgment of flow rate and color, which seriously restricts the automation and intelligent upgrading of filter press equipment.

[0024] To address these technical challenges in industrial filter presses, such as inclined and intersecting drainage pipes, minute and discontinuous water flow, difficulty in decoupling merged water flows, misaligned flow pipes, and inaccurate flow color judgment, this application proposes a method for detecting the flow rate of filter press filtrate. This method uses real-time images of a designated monitoring area of ​​the filter press as input. Based on a dual-domain core constraint mechanism of the spatial topological feature field of the drainage pipes and the potential field of the water flow-drainage pipe matching association, it transforms the discrete location, shape, and neighborhood topological relationships of the drainage pipes, as well as the location, area, and flow direction characteristics of the water flow, into continuous, quantifiable, and computable domain features. Combined with a pre-trained deep learning detection and segmentation network, it sequentially completes precise drainage pipe detection, precise water flow segmentation, one-to-one precise flow pipe matching, liquid discharge status recording, flow rate quantification calculation, and water flow color judgment, achieving fully automated, high-precision, stable, and reliable real-time monitoring of the filter press filtrate flow rate.

[0025] The advantages of this application are: 1. A method for accurate identification of intersecting drainage pipes was designed. By providing strong spatial constraints through the spatial topological feature field of drainage pipes, it can accurately distinguish drainage pipes in inclined, intersecting and overlapping states. The false detection rate and false detection rate are significantly reduced, the detection results are stable and unique, and it has real-time pose adaptation capability.

[0026] 2. A complete segmentation method for complex water flows was constructed. The method adopts a multi-color space fusion and merging water flow decoupling algorithm, which can achieve complete segmentation of small, discontinuous, transparent and merged water flows. The segmentation results are free of missing parts and false contours, effectively solving the segmentation problem of water flow and background convergence.

[0027] 3. A misalignment-free and precise matching logic for the flow pipe was designed. Based on the physical constraint matching of the associated potential field, the water flow and the drain pipe are strictly bound one-to-one, solving the problems of misalignment, one-to-many, and many-to-one matching. Invalid water flow interference is automatically eliminated, and the matching result conforms to the physical liquid discharge logic.

[0028] 4. It achieves precise flow quantification, and the three-dimensional color judgment mechanism is not affected by ambient light or dirt. It can stably distinguish the state of various filtrates, and the monitoring data can directly support the automated control of the filter press.

[0029] 5. It has high adaptability to industrial sites, lightweight process flow, integrated device design, strong hardware resistance to harsh environments, low deployment cost, stable operation, and can be adapted to various types of filter presses, effectively promoting the intelligent upgrading of solid-liquid separation equipment.

[0030] After introducing the application scenarios and design concepts of the embodiments of this application, the technical solutions provided by the embodiments of this application will be described below.

[0031] like Figure 1 As shown in the figure, this application provides a method for detecting the flow rate of filtrate in a filter press. The filter press includes a drain pipe, with each drain pipe corresponding to a flow of filtrate. The method includes the following steps: S101 acquires an RGB image of the filtrate monitoring area of ​​the filter press under test; S102 uses the drainage pipe detection model to process the RGB image and obtain the detection results of each drainage pipe; the drainage pipe outlet is set as the attraction potential source, and the attraction potential source intensity of each drainage pipe outlet is calculated based on the detection results. S103 uses a water flow segmentation network to process the RGB image to obtain the water flow morphology features and segmentation results of each filtrate flow; the centroid of each filtrate flow is set as the attracted potential point, and the attracted potential point features of each filtrate flow are calculated based on the segmentation results. S104 calculates the associated potential field value of each drain pipe and the filtrate flow based on the strength of the attraction potential source and the characteristics of the attraction potential point of the filtrate flow; and determines the drain pipe matched with each filtrate flow based on the associated potential field value. S105 determines the input characteristics of each filtrate flow based on the segmentation results, water flow morphology characteristics, correlation potential field values, and detection results; S106 inputs the input features into the water flow estimation model to obtain the flow rate of the filtrate from the filter press to be tested.

[0032] The method in this embodiment can achieve fully automated, high-precision, stable and reliable real-time monitoring of the filtrate flow rate of the filter press.

[0033] S101 acquires an RGB image of the filtrate monitoring area of ​​the filter press to be tested.

[0034] An RGB image is a digital image that uses red, green, and blue light mixed in different proportions to generate patterns. Each of the three color channels (red, green, and blue) is divided into 256 levels of brightness. In this step, an RGB image of the filtrate monitoring area of ​​the filter press to be tested is acquired. This image serves as the input to the method of this application, providing a basis for subsequent processing and analysis.

[0035] S102 uses the drainage pipe detection model to process the RGB image and obtain the detection results of each drainage pipe; the drainage pipe outlet is set as the attraction potential source, and the attraction potential source intensity of each drainage pipe outlet is calculated based on the detection results.

[0036] In this step, the drain pipe detection model is a pre-trained YOLOv11 model. The detection results include the detection box and number for each drain pipe. The detection process is as follows: Figure 2 As shown.

[0037] In some embodiments, the intensity of the attraction potential source at each drain pipe outlet is calculated based on the detection results, including: The attraction potential source strength of the i-th drain pipe is calculated according to the following formula. :

[0038] in, The two-dimensional coordinates of the outlet of the i-th drain pipe are determined by the detection frame; Let be the weight of the i-th drainage pipe; The L2 norm is used; specifically, a standard two-dimensional Cartesian coordinate system is established with the physical plane of the filtrate monitoring area below the filter plate of the filter press as the reference. Using the lower left corner of the monitoring area as the origin of the coordinate system, the horizontal direction to the right is the positive x-axis, and the vertical direction upward is the positive y-axis. This ensures that each pixel in the image is precisely mapped to the physical coordinate system, guaranteeing the physical consistency of subsequent field calculations and position determination.

[0039] Let be the three-dimensional field value at the outlet of the i-th drain pipe, expressed as:

[0040] in, The i-th drain pipe is located at coordinates The field strength of the topological characteristic field at a location represents the field component at that spatial location.

[0041] in, This represents the normalized value of the actual tilt angle of the i-th drainage pipe, reflecting the characteristics of the pipe's pose change. This is the normalized value of the actual length of the i-th drainage pipe, reflecting the morphological characteristics of the water pipe. It is the normalized value of the center-Euclidean distance between the i-th and j-th drainage pipes, reflecting the neighborhood topological relationship; It is the weighting coefficient of the outlet position of the i-th drain pipe, with the outlet area being given a higher weight. Represents coordinate points The normalized value of the Euclidean distance to the center of the source of the i-th drainage pipe; is the spatial attenuation coefficient of the electric field strength, which decreases exponentially with increasing distance; , , and These are the weighting coefficients; the field strength distribution diagram of the topological characteristic field is as follows: Figure 3 As shown.

[0042] For the i-th drain pipe at coordinates The field strength of the morphological characteristic field at a certain location characterizes the morphological features of the drainage pipe, such as its inclination and length.

[0043] in, Is the i-th drain pipe at coordinates The similarity of the contour shape at each location is calculated using the Hausdorff distance index, which is used to characterize the thickness of the tube body and the features of the edge contour. Let be the topological correlation field strength of the neighborhood of the i-th drainage pipe, representing the interleaving and overlapping relationship between the drainage pipe and its adjacent water pipes:

[0044] Where M represents the number of adjacent water pipes in the neighborhood of the i-th drainage pipe. The normalized value of the center distance between the i-th and j-th drainage pipes. The percentage of overlapping areas between adjacent drainage pipes represents the degree of overlap.

[0045] In this embodiment, the three-dimensional field data can continuously characterize the correlation between any pixel in the image and each drainage pipe. The field value changes smoothly with the physical distance to the field source, which can accurately distinguish different drainage pipes in tilted, intersecting, and overlapping states, providing strong prior space constraints for the subsequent drainage pipe detection network.

[0046] This embodiment generates a real-time drainage pipe feature field for the current frame image. By analyzing the field overlap and gradient change rate of the feature field, it accurately identifies the boundary contours of intersecting drainage pipes, ensuring a high degree of match between the detection results and the field constraints. This fusion update mechanism can adapt in real time to minor positional shifts and vibration deformations of drainage pipes after long-term operation, enabling the detection process to have stable spatial self-adaptation capabilities and providing a unique, stable, and accurate basis for drainage pipe positioning in subsequent water flow matching processes.

[0047] S103 uses a water flow segmentation network to process the RGB image to obtain the water flow morphology features and segmentation results of each filtrate flow; the centroid of each filtrate flow is set as the attracted potential point, and the attracted potential point features of each filtrate flow are calculated based on the segmentation results.

[0048] In this step, the actual physical area of ​​the filtrate flow is represented by a flow segmentation network. The aspect ratio of the water flow profile and the perimeter of the water flow segmentation profile together constitute the water flow morphological characteristics.

[0049] In this step, the segmentation result for each filtrate flow includes: a binarized segmentation mask and a number for the filtrate flow. The flow segmentation network uses a Segformer network.

[0050] In some embodiments, the attraction potential point characteristics of each filtrate flow are calculated based on the segmentation results, including: The total number of pixels in the foreground region of the binarized segmentation mask of the j-th filtrate flow is counted to obtain the pixel area of ​​the j-th filtrate flow. The pixel area of ​​the j-th stream of filtrate is mapped to the actual physical area using the camera's physical calibration scale. ; The inter-frame optical flow method, Lucas-Kanade, was used to calculate the velocity vector of the j-th filtrate stream. ; right and Normalization is performed to obtain the normalized actual physical area. and normalized water flow velocity vector ; The attraction potential characteristic of the j-th stream of filtrate is calculated using the following formula. :

[0051] in, It is an L2 norm.

[0052] S104 calculates the associated potential field value of each drain pipe and the filtrate flow based on the strength of the attraction potential source and the characteristics of the attraction potential point of the filtrate flow; based on the associated potential field value, it determines the drain pipe matched to each filtrate flow.

[0053] In some embodiments, based on the strength of the attraction potential source and the characteristics of the attraction potential point of the filtrate flow, the associated potential field value of each drain pipe and the filtrate flow is calculated, including: The associated potential field value of the i-th drain pipe and the j-th filtrate flow is calculated according to the following formula. :

[0054] in, Let be the physical Euclidean distance between the outlet of the i-th drain pipe and the j-th stream of filtrate, determined by the position of the outlet of the i-th drain pipe. The centroid coordinates of the j-th filtrate flow Calculated; Let be the angle between the direction of liquid discharge from the i-th drain pipe and the direction of water flow from the j-th stream. All are weighting coefficients.

[0055] This embodiment sets a fixed potential field range for each attractive potential source, automatically eliminating invalid matching interference across regions and over long distances, ensuring that the matching results conform to physical logic; the distribution of the associated potential field values ​​is as follows: Figure 4 As shown.

[0056] In some embodiments, determining the drain pipe matched to each filtrate flow based on the associated potential field value of each drain pipe and the filtrate flow includes: The drain pipe corresponding to the maximum value of the associated potential field of the j-th filtrate flow and each drain pipe is determined as the drain pipe matched with the j-th filtrate flow.

[0057] This embodiment sets a fixed potential field range for each attractive potential source, automatically eliminating invalid matching interference across regions and over long distances, and ensuring that the matching results conform to physical logic.

[0058] In some embodiments, the method further includes: marking successfully matched drain pipes as liquid discharge states, marking unmatched drain pipes as liquid-free states, and automatically determining water flows without corresponding matched drain pipes as invalid water flows.

[0059] S105 determines the input characteristics of each filtrate flow based on the segmentation results, flow morphology characteristics, correlation potential field values, and detection results.

[0060] In some embodiments, the input characteristics of each filtrate flow are determined based on the segmentation results of each filtrate flow, the flow morphology characteristics, the detection results of the matched drain pipe, and the associated potential field value between each filtrate flow and the matched drain pipe; including: The actual physical area of ​​the j-th filtrate flow The water flow morphology features, composed of the aspect ratio of the water flow profile and the perimeter of the water flow segmentation profile, the detection frame size of the matched drain pipe, and the potential field value associated with the j-th filtrate flow and the matched drain pipe are dimensionally spliced ​​to obtain the input features of the j-th filtrate flow.

[0061] S106 inputs the input features into the water flow estimation model to obtain the flow rate of the filtrate from the filter press to be tested.

[0062] In this step, the flow estimation model consists of multiple sequentially connected fully connected layers.

[0063] In some embodiments, it also includes: The flow region is obtained by dividing the foreground region of the binarized segmentation mask based on the j-th filtrate flow; The K-means clustering algorithm is used to calculate the main color cluster center of the water flow region to determine the color of the j-th filtrate flow. Based on a preset color threshold, the color of the j-th filtrate stream is used to determine the flow state, including: normal filtrate, filtrate containing impurities, and filtrate containing contaminants.

[0064] Based on the same inventive concept, this application provides a filter press filtrate flow rate detection device, see reference. Figure 5 As shown, the filter press filtrate flow rate detection device 200 provided in this application embodiment includes at least: The acquisition unit 201 is used to acquire an RGB image of the filtrate monitoring area of ​​the filter press to be tested; The first processing unit 202 is used to process the RGB image using the drainage pipe detection model to obtain the detection results of each drainage pipe; the drainage pipe outlet is set as the attraction potential source, and the attraction potential source intensity of each drainage pipe outlet is calculated based on the detection results. The second processing unit 203 is used to process the RGB image using a water flow segmentation network to obtain the water flow morphology features and segmentation results of each filtrate water flow; the centroid of each filtrate water flow is set as the attracted potential point, and the attracted potential point features of each filtrate water flow are calculated based on the segmentation results. Matching unit 204 is used to calculate the associated potential field value of each drain pipe and the filtrate flow based on the strength of the attraction potential source and the characteristics of the attraction potential point of the filtrate flow; and to determine the drain pipe matched to each filtrate flow based on the associated potential field value. The determination unit 205 is used to determine the input characteristics of each filtrate flow based on the segmentation results, water flow morphology characteristics, correlation potential field values ​​and detection results; The detection unit 206 is used to input the input features into the water flow estimation model to obtain the flow rate of the filtrate from the filter press to be detected.

[0065] It should be noted that the principle of the filter press filtrate flow rate detection device 200 provided in this application embodiment to solve the technical problem is similar to the method provided in this application embodiment. Therefore, the implementation of the filter press filtrate flow rate detection device 200 provided in this application embodiment can refer to the implementation of the method provided in this application embodiment, and the repeated parts will not be described again.

[0066] Based on the same inventive concept, embodiments of this application also provide an electronic device, such as... Figure 6 As shown, it includes a memory and a processor. The memory stores an executable program, and the processor executes the executable program to implement the steps of the filter press filtrate flow rate detection method provided in the above embodiment.

[0067] The aforementioned processor can be a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The aforementioned PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0068] Since the electronic device described in this application embodiment is an electronic device equipped with a memory for implementing the filter press filtrate flow rate detection method disclosed in this application embodiment, those skilled in the art can understand the structure and variations of the electronic device described in this application embodiment based on the filter press filtrate flow rate detection method described in this application embodiment, and therefore will not be described again here.

[0069] This application also provides a computer-readable storage medium storing a computer program thereon, which, when run by a processor, implements the steps of the filter press filtrate flow rate detection method provided in the above embodiments.

[0070] The storage medium in this embodiment may be included in an electronic device; or it may exist independently and not be assembled into an electronic device. The storage medium carries one or more computer programs, which, when executed, implement the steps of the filter press filtrate flow rate detection method provided in the above embodiment.

[0071] It should be understood that the various solutions in this embodiment have the same technical effects as those in the above method embodiments, and will not be repeated here.

[0072] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. Optionally, specific examples in this embodiment can refer to the examples described in any embodiment of this application, which will not be repeated here. Obviously, those skilled in the art should understand that the various modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular hardware and software combination.

[0073] This application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the filter press filtrate flow rate detection method provided in the above embodiments.

[0074] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions targeted in the blocks may occur in a different order than those targeted in the drawings. For example, two consecutively represented blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0075] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. Multitasking and parallel processing may be advantageous in certain environments. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

Claims

1. A method for detecting the flow rate of filtrate in a filter press, wherein the filter press includes a drain pipe, each drain pipe corresponding to a flow of filtrate, characterized in that, include: Acquire RGB images of the filtrate monitoring area of ​​the filter press under test; The RGB image is processed using a drainage pipe detection model to obtain the detection results for each drainage pipe. The outlet of the drain pipe is set as the attraction potential source, and the attraction potential source intensity of each drain pipe outlet is calculated based on the detection results. The RGB image is processed using a water flow segmentation network to obtain the water flow morphology features and segmentation results of each filtrate flow; the centroid of each filtrate flow is set as the attracted potential point, and the attracted potential point features of each filtrate flow are calculated based on the segmentation results. Based on the strength of the attraction potential source and the characteristics of the attraction potential point of the filtrate flow, the associated potential field value of each drain pipe and the filtrate flow is calculated. Based on the associated potential field value, determine the drain pipe that matches each of the filtrate water flows; The input characteristics of each filtrate flow are determined based on the segmentation results, the flow morphology characteristics, the correlation potential field value, and the detection results. The input features are input into the water flow estimation model to obtain the flow rate of the filtrate from the filter press to be tested.

2. The method according to claim 1, characterized in that, Based on the detection results, the attraction potential source intensity of each of the drainage pipe outlets is calculated, including: The two-dimensional coordinates of each drainage pipe are determined based on the drainage pipe detection frame in the detection results; The attraction potential source intensity of each of the drainage pipe outlets is calculated based on the two-dimensional coordinates.

3. The method according to claim 1, characterized in that, Based on the segmentation results, the attraction potential point characteristics of each of the filtrate flows are calculated, including: The pixel area of ​​each filtrate flow is determined based on the binarized segmentation mask in each segmentation result; The attraction potential point features of each filtrate flow are calculated based on the pixel area.

4. The method according to claim 1, characterized in that, The RGB images are processed using a water flow segmentation network to obtain the water flow morphology features of each of the filtrate water flows, including: The RGB image is processed using a water flow segmentation network to obtain the actual physical area of ​​each filtrate water flow, the aspect ratio of the water flow profile, and the perimeter of the water flow segmentation profile. The actual physical area, the aspect ratio of the water flow profile, and the perimeter of the water flow segmentation profile are used to construct the water flow morphology characteristics of each filtrate flow.

5. The method according to claim 1, characterized in that, Based on the associated potential field value, the drain pipes matching the filtrate flow are determined, including: Corresponding to one of the filtrate water flows, the associated potential field value of each of the drain pipes is matched; The drain pipe corresponding to the maximum value among the associated potential field values ​​is determined as the drain pipe of the filtrate water flow.

6. The method according to claim 1, characterized in that, The method further includes: The successfully matched drain pipe is marked as being in the liquid discharge state, the drain pipe that is not matched with the filtrate flow is marked as being in the liquid-free state, and the flow without a corresponding matching drain pipe is automatically determined as an invalid flow.

7. The method according to claim 1, characterized in that, Based on the segmentation results, the water flow morphology features, the correlated potential field value, and the detection results, the input features of each of the filtrate water flows are determined, including: The input features of each of the filtrate water flows are obtained by dimensionally concatenating the water flow morphology features, the drainage pipe detection box in the segmentation result, and the associated potential field value.

8. A device for detecting the flow rate of filtrate in a filter press, characterized in that, include: The acquisition unit is used to acquire RGB images of the filtrate monitoring area of ​​the filter press to be tested; The first processing unit is used to process the RGB image using the drain pipe detection model to obtain the detection results of each drain pipe; The outlet of the drain pipe is set as the attraction potential source, and the attraction potential source intensity of each drain pipe outlet is calculated based on the detection results. The second processing unit is used to process the RGB image using a water flow segmentation network to obtain the water flow morphology features and segmentation results of each of the filtrate water flows; The centroid of each of the filtrate streams is set as the point of attraction potential, and the characteristics of the point of attraction potential of each of the filtrate streams are calculated based on the segmentation results. A matching unit is used to calculate the associated potential field value of each drain pipe and the filtrate flow based on the strength of the attraction potential source and the attraction potential point characteristics of the filtrate flow. Based on the associated potential field value, determine the drain pipe that matches each of the filtrate water flows; The determining unit is used to determine the input features of each of the filtrate flows based on the segmentation results, the water flow morphology features, the correlation potential field value, and the detection results; The detection unit is used to input the input features into the water flow estimation model to obtain the flow rate of the filtrate from the filter press to be detected.

9. An electronic device, characterized in that, include: processor; And a memory storing a program, wherein the program includes instructions that, when executed by the processor, cause the processor to perform the method as described in any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method as described in any one of claims 1-7.

Images