Self-identifying intelligent cleaning method, device, medium and program product for closed space

Abstract

This invention discloses a self-identifying intelligent cleaning method, device, medium, and program product for enclosed spaces, relating to the field of image data recognition technology. The method includes: acquiring a material image; identifying the material image based on computer vision technology to obtain material characteristic information of easily clogged areas and the flow rate of inlets and outlets, the material characteristic information including at least one of adhesion degree, color information, and texture information; determining whether the easily clogged area is blocked based on the material characteristic information and flow rate; if so, determining a target cleaning plan based on the material characteristic information; generating a first cleaning control command based on the target cleaning plan; and sending the first cleaning control command to the cleaning device. This invention uses computer vision technology to identify material characteristic information, achieving accurate blockage judgment and target cleaning plan formulation, reducing damage to equipment during cleaning, and improving the intelligence level, efficiency, and safety of cleaning operations.

Description

Technical Field

[0001] This application relates to the field of image data recognition technology, and in particular to a self-identifying intelligent cleaning method, device, medium, and program product for enclosed spaces. Background Technology

[0002] Feed troughs are common equipment in industrial production, used to transport various bulk materials. However, because the feed trough is a relatively enclosed space, moisture from the material is not easily released, which can easily lead to excessively high material humidity. This causes the material to stick together inside the trough, making it prone to blockage. This not only affects production efficiency but may also damage related equipment.

[0003] Manual cleaning is difficult and dangerous. Therefore, the usual technical approach is to install air cannons or vibration devices at areas prone to blockage, using these devices to clear the blockage when it occurs.

[0004] However, air cannons or vibration devices exert a large force on the outside when they are working, which can easily damage production equipment. In addition, their cleaning and unblocking effects are relatively limited, and blockages in easily clogged areas are difficult to remove completely. Summary of the Invention

[0005] To address the aforementioned technical problems and deficiencies, the purpose of this invention is to provide a self-identifying intelligent cleaning method, device, medium, and program product for enclosed spaces. This product can automatically identify and judge the material characteristics and flow rate of easily clogged parts of the feed trough using computer vision technology, thereby achieving accurate blockage judgment and target cleaning plan formulation. It eliminates the need for air cannons or vibration devices to clear blockages, reducing damage to production equipment during cleaning and improving the level of intelligence, efficiency, and safety of cleaning operations.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a self-identifying intelligent cleaning method for enclosed spaces, applied to a control device of a cleaning system. The cleaning system further includes a cleaning device and a camera. The cleaning device is installed at a clog-prone section of the feeding trough, and the camera's shooting direction is towards the clog-prone section and the material inlet / outlet. The cleaning device and the camera are respectively communicatively connected to the control device. The method includes: acquiring material images of the clog-prone section and the material inlet / outlet, the material images being captured by the camera; identifying the material images based on computer and vision technology to obtain material characteristic information of the clog-prone section and the material inlet / outlet flow rate, the material characteristic information including at least one of adhesion degree, color information, and texture information; determining whether the clog-prone section is in a blocked state based on the material characteristic information and flow rate; if so, determining a target cleaning scheme based on the material characteristic information and flow rate; generating a first cleaning control command based on the target cleaning scheme, the first cleaning control command being used to control the cleaning device to perform cleaning operations to achieve the target cleaning scheme; and sending the first cleaning control command to the cleaning device.

[0007] By employing the above embodiments and acquiring material images of easily clogged areas and analyzing them using computer vision technology, this solution achieves accurate identification of information such as the degree of adhesion, flow rate, color information, and texture information of materials in the enclosed space of the feeding trough. This intelligent identification system can monitor the material status in real time and promptly detect potential blockage problems. Combining this material information, the system can accurately determine whether there is a risk of blockage at easily clogged areas, thereby taking preventive measures. When a blockage is detected, the system automatically formulates a suitable cleaning plan and generates corresponding control commands. These commands guide the cleaning device to perform effective cleaning operations, ensuring that blockages are cleared in a timely manner and guaranteeing the smooth flow of production. The entire process is automated and intelligent, significantly improving cleaning efficiency and accuracy, reducing manual intervention, lowering production costs, eliminating the need for air cannons or vibration devices to clear blockages, reducing damage to production equipment during cleaning, realizing intelligent maintenance and management of the feeding trough, and improving the operational safety and reliability of the entire system.

[0008] In some embodiments, the cleaning system further includes a lidar, and the material feature information further includes the three-dimensional shape information of the material. Before the step of determining whether the easily clogged part is in a clogged state based on the material feature information, the system further includes: acquiring the three-dimensional shape information of the material at the easily clogged part through the lidar, and the three-dimensional shape information is used to determine whether there are foreign objects or large pieces of material at the easily clogged part.

[0009] By using cameras to simultaneously identify adhesion levels, flow rates, color information, and texture information, combined with precise 3D shape information acquired by LiDAR, the control device can achieve comprehensive monitoring and in-depth analysis of material conditions. This multi-dimensional data fusion method greatly improves the accuracy of identifying material flow and potential blockages, enabling timely detection and response to complex situations such as adhesion, foreign object intrusion, and large material accumulation. Intelligent analysis algorithms can adjust cleaning strategies based on real-time data, optimizing cleaning paths and operating modes to effectively prevent and resolve blockages, ensuring continuous production line operation and efficiency. This also reduces the need for manual inspection and intervention, improving operational safety and intelligence.

[0010] In some embodiments, the step of determining whether a clogged area is in a blocked state based on material characteristic information and flow rate includes: determining the adhesion index B, flow rate index Q, color value C, and texture entropy value T based on the degree of adhesion, flow rate, color information, and texture information, respectively; and substituting the adhesion index B, flow rate index Q, color value C, and texture entropy value T into the blockage formula to calculate the blockage index DI. The mathematical expression of the blockage formula is as follows:

[0011] ;

[0012] Among them, B thr B represents the adhesion threshold. max Q represents the maximum possible value of the adhesion index B. nor This represents the theoretical normal flow rate of the easily clogged section under unobstructed conditions, where K is the material flow rate attenuation coefficient, and C... thr C represents the color threshold, indicating the critical point at which color changes. nor The normal color of the material, T thr T represents the texture entropy threshold, indicating the critical point at which the texture entropy value changes. nor The normal texture entropy value of the material is represented by w1, w2, w3, and w4, which are weighting coefficients. The blockage index DI is used to determine whether the blockage-prone parts are in a blocked state.

[0013] Using the above embodiments, based on adhesion degree, flow rate, color information, and texture information, combined with a clogging formula, the clogging status of easily clogged areas can be accurately determined. It can monitor and accurately assess clogging risks in real time, promptly triggering the cleaning device to perform necessary cleaning operations, thereby effectively preventing and resolving clogging problems. This not only improves production efficiency and reduces downtime but also lowers maintenance costs and enhances the stability and reliability of the feed trough. Through intelligent clogging prediction and response, this solution significantly improves the overall performance of the industrial production process.

[0014] In some embodiments, the cleaning system further includes a humidity sensor, which is located at a clog-prone area. After determining whether the clog-prone area is blocked based on material characteristics and flow rate, the system further includes: if the clog-prone area is not blocked, acquiring the material humidity information of the clog-prone area, which is obtained by the humidity sensor; and when the material humidity information is greater than a set humidity threshold, sending a second cleaning control command to the cleaning device, which is used to control the cleaning device to perform a preset cleaning operation.

[0015] In the above embodiment, the control device continuously monitors the material moisture information at easily clogged areas. This information is collected in real time by moisture sensors installed at key locations. Once the detected material moisture exceeds a preset moisture threshold, indicating that the material may be too wet and pose a high risk of clogging, the control device will respond immediately. At this time, the control device will send a second cleaning control command to the cleaning device. This command contains a series of operating parameters to guide the cleaning device to perform specific cleaning operations. Preset cleaning operations may include increasing cleaning intensity, adjusting cleaning frequency, changing cleaning mode, or using a special cleaning program to more effectively handle wet materials and prevent clogging. After receiving the second cleaning control command, the cleaning device will automatically adjust its working state and perform the required cleaning tasks. This series of actions aims to reduce material moisture in a timely manner, eliminate potential clogging hazards, and ensure the continuity and efficiency of the production process. Through this intelligent monitoring and response mechanism, the system can proactively prevent and resolve potential clogging problems, reducing production interruptions and maintenance costs.

[0016] In some embodiments, the cleaning system further includes a flow sensor, which is located at a clog-prone area. After determining whether the clog-prone area is blocked based on material characteristics and flow rate, the system further includes: if the clog-prone area is not blocked, acquiring material flow rate information of the clog-prone area, which is obtained by the flow sensor; and when the material flow rate information is less than a set flow rate threshold, sending a third cleaning control command to the cleaning device, which is used to control the cleaning device to perform a preset cleaning operation.

[0017] In the above embodiment, the control device connects to a flow sensor installed in the feeding trough system to acquire real-time material flow information at easily clogged sections. The flow sensor monitors and measures the amount of material passing through a specific cross-section per unit time, providing crucial flow data to the control device. If the control device detects that the material flow is below a set flow threshold, this may indicate that material flow is obstructed or there is a risk of blockage. In this case, the control device sends a third cleaning control command to the cleaning device. This command is based on a predefined cleaning strategy and operating procedure, designed to guide the cleaning device to perform targeted cleaning operations. These operations may include adjusting the cleaning device's trajectory, cleaning intensity, or activating a specific cleaning mode to effectively remove potential blockages and restore normal material flow. Upon receiving the third cleaning control command, the cleaning device automatically executes the corresponding cleaning tasks, such as activating the brushes, adjusting the scraper position, or activating the vacuum system, to ensure that easily clogged sections are cleaned promptly, preventing further blockages and thus ensuring the continuity and efficiency of the entire production process. Through this automated response mechanism, the system can quickly identify and resolve potential blockage problems, reducing the risk of production interruptions.

[0018] In some embodiments, after sending the first cleaning control command to the cleaning device, the method further includes: after receiving the cleaning completion information sent by the cleaning device, acquiring an image of the cleaned material captured by a camera; determining whether the cleaning completion result of the cleaning device is qualified based on the image of the cleaned material; if not, adjusting the target cleaning plan based on the image of the cleaned material.

[0019] In the above embodiment, after the cleaning device completes its cleaning task and sends cleaning completion information to the control device, the control device immediately activates a camera to capture images of the cleaned material. These images are then sent to the control device for analysis to evaluate the cleaning effect. The control device uses computer vision technology to perform a detailed analysis of the cleaned material images, identifying the material distribution, adhesion degree, and the presence of residual blockages. By comparing the differences between the images before and after cleaning, the control device can determine whether the cleaning result achieved the predetermined standard. If the analysis results show that the cleaning result is unqualified, i.e., there are still material residues or blockages that have not been completely removed, the control device will adjust the target cleaning plan based on the specific situation in the cleaned material images. This may include increasing the cleaning intensity, changing the operating mode of the cleaning tools, replanning the cleaning path, or extending the cleaning time. The adjusted target cleaning plan will then be converted into specific control instructions, which the control device will resend to the cleaning device to guide it to perform secondary cleaning or other necessary operations until the cleaning result meets the requirements. This process ensures the efficiency and thoroughness of the cleaning task, guarantees the normal operation of the feeding trough, and the continuity of the production process.

[0020] In some embodiments, after determining whether a blockage-prone part is in a blocked state based on material information and flow rate, the method includes: if it is not in a blocked state, determining the estimated blockage time of the blockage-prone part based on the current material humidity information and material flow rate information of the blockage-prone part; determining the cleaning task execution time based on the estimated blockage time; and sending a first cleaning control command to the cleaning device when the cleaning task execution time is reached.

[0021] By analyzing and utilizing the material moisture and flow information at easily clogged locations using the above embodiments, this invention can accurately predict the timing of blockages and thus plan the execution time of cleaning tasks. This prediction-based cleaning strategy can effectively prevent blockages, improve the initiative of the cleaning system, and ensure the continuity and stability of the production process. Simultaneously, this method also helps optimize the allocation of cleaning resources, reduce unnecessary cleaning operations, lower maintenance costs, and improve the overall efficiency and reliability of cleaning operations.

[0022] In a second aspect, embodiments of this application provide a control device comprising: one or more processors and a memory; the memory is coupled to the one or more processors and is used to store computer program code, the computer program code including computer instructions, wherein the one or more processors invoke the computer instructions to cause the control device to perform the method described in the first aspect and any possible implementation thereof.

[0023] Thirdly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a control device, cause the control device to perform the method described in the first aspect and any possible implementation thereof.

[0024] Fourthly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on a control device, cause the control device to perform the method described in the first aspect and any possible implementation thereof.

[0025] It is understood that the control device provided in the second aspect, the storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0026] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0027] 1. By acquiring material images of easily clogged areas and analyzing them using computer vision technology, this solution achieves accurate identification of material adhesion, flow rate, color information, and texture information. This intelligent identification system can monitor the material status in real time and promptly detect potential blockage problems. Combining this material information, the system can accurately determine whether there is a risk of blockage at easily clogged areas, thereby taking preventive measures. When a blockage is detected, the system automatically formulates a suitable cleaning plan and generates corresponding control commands. These commands guide the cleaning device to perform effective cleaning operations, ensuring that blockages are cleared in a timely manner and guaranteeing smooth production processes. The entire process is automated and intelligent, significantly improving cleaning efficiency and accuracy, reducing manual intervention, lowering production costs, eliminating the need for air cannons or vibration devices for unblocking, reducing damage to production equipment during cleaning, realizing intelligent maintenance and management of the feed trough, and improving the operational safety and reliability of the entire system.

[0028] 2. Based on adhesion level, flow rate, color information, and texture information, combined with clogging formulas, the clogging status of easily clogged areas can be accurately determined. It can monitor and accurately assess clogging risks in real time, promptly triggering the cleaning device to perform necessary cleaning operations, thereby effectively preventing and resolving clogging problems. This not only improves production efficiency and reduces downtime but also lowers maintenance costs and enhances system stability and reliability. Through intelligent clogging prediction and response, this solution significantly improves the overall performance of industrial production processes.

[0029] 3. After the cleaning device completes its cleaning task and sends a completion message to the control device, the control device immediately activates a camera to capture images of the cleaned material. These images are then sent to the control device for analysis to evaluate the cleaning effect. The control device uses computer vision technology to perform a detailed analysis of the cleaned material images, identifying the material distribution, adhesion degree, and the presence of residual blockages. By comparing the differences between the images before and after cleaning, the control device can determine whether the cleaning results have met the predetermined standards. If the analysis results show that the cleaning results are unqualified, i.e., there are still material residues or blockages that have not been completely removed, the control device will adjust the target cleaning plan based on the specific situation in the cleaned material images. This may include increasing the cleaning intensity, changing the operating mode of the cleaning tools, replanning the cleaning path, or extending the cleaning time. The adjusted target cleaning plan will then be converted into specific control instructions, which the control device will resend to the cleaning device to guide it to perform secondary cleaning or other necessary operations until the cleaning results meet the requirements. This process ensures the efficiency and thoroughness of the cleaning task, guarantees the normal operation of the feeding trough, and the continuity of the production process. Attached Figure Description

[0030] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention. It is obvious that the drawings described below are merely some embodiments of the invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings:

[0031] Figure 1 This is a schematic diagram of the working scene of the cleaning system according to an embodiment of the present invention;

[0032] Figure 2 This is a side view of the cleaning device in an embodiment of the present invention;

[0033] Figure 3 This is a top view of the cleaning device in an embodiment of the present invention;

[0034] Figure 4 This is a flowchart illustrating a self-identifying intelligent cleaning method for enclosed spaces according to an embodiment of the present invention.

[0035] Figure 5 This is a flowchart illustrating another method for self-identifying intelligent cleaning in enclosed spaces according to an embodiment of the present invention.

[0036] Figure 6 This is a schematic diagram of the architecture of an electronic device in an embodiment of this application. Detailed Implementation

[0037] The terminology used in the following embodiments of the present invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the invention, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in the invention refers to any or all possible combinations comprising one or more of the listed items.

[0038] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0039] This invention provides a cleaning system, which includes a control device, a cleaning device, and a camera, such as... Figure 1As shown, the cleaning device is arranged at the easily blocked part of the feeding trough, the shooting direction of the camera faces the easily blocked part and the material inlet and outlet of the feeding trough, and the cleaning device and the camera are respectively connected to the control device in communication.

[0040] Among them, the control device is the core of the entire cleaning system and can be an intelligent device integrating advanced processors and software algorithms. It is responsible for receiving the image data from the camera and performing image analysis based on computer vision technology to identify and evaluate the characteristic information of the material. The control device is also responsible for formulating a cleaning strategy according to the analysis results and generating corresponding cleaning control instructions. In addition, the control device manages the communication with the cleaning device to ensure the accurate transmission and execution of the instructions. It can be connected to a wider industrial network to achieve remote monitoring and data integration, and improve the intelligent level of the entire system.

[0041] The cleaning device can be a cleaning robot or a robotic arm, specifically including a driving motor and a sweeper. A slide rail is installed at the easily blocked part of the feeding trough, the sweeper is installed on the slide rail, and the driving motor can be arranged outside the feeding trough. The driving motor drives the sweeper to move on the slide rail, thereby cleaning the material at the easily blocked part and dredging the feeding trough. As Figure 2 and 3 shown, the structure of the sweeper can be in the shape of a "king", with multiple cleaning arms, and this shape can improve the cleaning efficiency.

[0042] The camera plays the role of "visual perception" in the cleaning system and is responsible for capturing real-time images or videos of the easily blocked part of the feeding trough. These image data are crucial for the control device to analyze the characteristic information of the material. The camera usually needs to have high resolution and good imaging performance to ensure clear images can be obtained under various lighting conditions. In addition, the camera may also need to have the characteristics of waterproof, dustproof and anti-vibration to adapt to the harsh industrial environment. To achieve the best monitoring effect, the installation position and angle of the camera need to be carefully designed to cover the key areas of the easily blocked part.

[0043] The present invention also provides a self-identifying intelligent cleaning method for a closed space, which is applied to the cleaning system and is executed by the control device. As Figure 4 shown, the method includes the following steps:

[0044] Step 101, obtain the material images of the easily blocked part and the material inlet and outlet of the feeding trough.

[0045] Among them, the material images are obtained by the camera shooting. In this embodiment, the feeding trough can be regarded as a closed space for storing and transporting materials.

[0046] Specifically, the control device communicates with a camera, which captures real-time images of materials at clog-prone areas and material inlets / outlets. It automatically takes and records these images. These images are then transmitted to the control device via wired or wireless network. The control device receives and processes these images, using computer vision technology to analyze the material's characteristics, thereby determining the blockage status and formulating a cleaning strategy. This process ensures that the control device acquires the latest and most accurate material images, providing a reliable visual basis for intelligent cleaning.

[0047] The material images can be a series of continuously captured still images or a series of continuous dynamic video frames that constitute a video. In this solution, the control device acquires material images of easily clogged areas in the following two ways:

[0048] Images captured in succession:

[0049] In this method, the camera is set to continuous shooting mode, periodically or as needed capturing still images of areas prone to blockage. These images can be captured at set intervals, such as automatically taking pictures at regular intervals. The control unit then receives these still images from the camera and analyzes them.

[0050] Video frames:

[0051] Another approach is to use a camera to record video, and the control device extracts specific frames from the recorded video stream as material images. This method provides richer information because it includes not only static information but also captures the dynamic changes in material flow. The control device can set the video frame rate to suit different monitoring needs and analytical precision.

[0052] Whether using continuously captured images or video frames, the control device needs to possess efficient image processing capabilities to extract key information about the material's characteristics. Dynamic changes in the material images allow observation of the rate and quantity of material passing through the inlet and outlet, thus reflecting the magnitude of the material flow.

[0053] Material images of easily clogged areas can reflect material characteristics such as adhesion, color, and texture. Material images of inlets and outlets can reflect material flow rates.

[0054] Step 102: Based on computer and vision technology, identify material images to obtain material characteristic information of easily blocked parts and the flow rate of material inlet and outlet.

[0055] The material characteristic information includes at least one of the following: adhesion degree, color information, and texture information.

[0056] The degree of adhesion refers to the bonded state of materials at their contact surfaces due to the adhesive forces between them. In industrial production, materials may adhere due to humidity, chemical properties, or other factors, which can lead to malfunctions or complete blockages in equipment. Quantifying the degree of adhesion can help determine whether materials are prone to aggregation or blockage, allowing for preventative or cleaning measures to be taken.

[0057] Flow rate refers to the speed and quantity of material passing through the feed chute. On a production line, normal flow rate indicates smooth material flow, while an abnormal decrease or instability in flow rate may indicate a blockage.

[0058] Color information refers to the characteristic color of a material's surface, reflecting attributes such as its chemical composition, humidity, temperature, and localized stress. Color information extraction is typically achieved through image processing techniques, such as color histograms and color moments. Changes in color may indicate alterations in the material's state; for example, mutual compression between materials can cause localized color changes, or high humidity can lead to significant color variations, both of which may be associated with clogging risk. By analyzing color information, the state of materials can be monitored and assessed.

[0059] Texture information describes the visual features of recurring patterns in a material image, reflecting the structure and composition of the material surface. Texture information can be extracted using techniques such as Gray-Level Co-occurrence Matrix (GLCM), Gabor filters, and Local Binary Pattern (LBP). Variations in texture may indicate material accumulation, compression, particle size variations, or other factors that could lead to blockage. Quantifying texture information provides a better understanding of the material's physical state and flow characteristics.

[0060] In this step, the control device uses advanced computer vision technology to perform in-depth analysis of the material images acquired from the camera in order to identify and extract the degree of adhesion, flow rate, color information, and texture information.

[0061] First, the control unit preprocesses the image, including adjusting contrast, brightness, and color balance, and applying denoising algorithms to eliminate random noise and improve image quality. This step is crucial for subsequent image analysis because it ensures that the features extracted from the image are accurate and reliable.

[0062] Next, the control device employs image segmentation techniques to separate the material from the background. This is typically achieved through methods such as color thresholding, texture analysis, or edge detection. These techniques identify the boundaries and shapes of the material in the image, providing a basis for assessing the degree of adhesion. Then, the control device uses morphological operations, such as erosion and dilation, to further refine the morphological features of the material in order to more accurately identify the adhered areas and separated particles.

[0063] To assess the degree of material adhesion, the control device calculates a series of image features, such as region area, perimeter, convex hull, fill power, and texture features. These features reflect the degree of dispersion and tightness of the material. For example, a region with high fill power may indicate that the material is tightly bonded, while a region with complex texture features may indicate that there are many gaps between the materials.

[0064] Simultaneously, the control unit assesses flow rate by analyzing material flow patterns in continuous image sequences or video frames. This involves object tracking algorithms, such as optical flow or Kalman filters, to track the trajectories of material particles in consecutive frames. By measuring the velocity and direction of these trajectories, the control unit can estimate the average flow velocity and flow rate of the material. Furthermore, the control unit can analyze patterns in material flow, such as the presence of congestion points or flow interruptions, which are crucial information for identifying potential blockages.

[0065] Color information can be extracted from various color spaces, such as RGB (Red, Green, Blue), HSV (Hue, Saturation, Brightness), and LAB. In computer vision, images are typically converted from their original color space to a more suitable one for analysis, and then color histograms or other color descriptors are calculated to quantify color features. For example:

[0066] Color histogram: Statistical analysis of the frequency of each color in an image, which can be used to analyze color distribution and variation.

[0067] Color moments: Characteristic values ​​that describe the color distribution, including the color mean, variance, etc.

[0068] Color correlation diagram: Analyzes the relationships and patterns between colors in a color space.

[0069] Texture information can be extracted using various texture analysis methods. For example:

[0070] Gray-level co-occurrence matrix (GLCM): Extracts texture features by analyzing the spatial relationships between pixels.

[0071] Gabor filters: filters with different frequencies and directions are used to capture texture information in an image.

[0072] Local Binary Pattern (LBP): Extracts local texture patterns by comparing the brightness of a pixel with that of its neighbors.

[0073] Wavelet transform: a multi-scale analysis method that can capture texture information at different frequencies.

[0074] Step 103: Based on the material characteristics and flow rate, determine whether the blockage-prone parts are in a blocked state.

[0075] If it is determined that there is no blockage, return to step 101 and continue monitoring. If it is determined that there is a blockage, proceed to step 104.

[0076] Specifically, the control device determines whether the easily clogged parts of the feed trough are blocked by comprehensively analyzing the material characteristic information and flow rate extracted from the material images obtained from the camera.

[0077] The degree of adhesion is assessed by calculating the morphological characteristics of the material in the image, such as the degree of aggregation between particles, surface coverage, and texture uniformity. High adhesion is typically characterized by tight bonding between material particles and large areas of continuous coverage, which may lead to flow obstruction.

[0078] The flow rate of materials at the inlet and outlet is determined by tracking changes in the material's position in a continuous sequence of images. The control device measures the amount of material passing through the inlet and outlet areas per unit time, as well as the speed and continuity of the material flow. A significant decrease in flow rate or an abnormal interruption in the flow pattern may indicate that the material flow is obstructed.

[0079] Under normal circumstances, the color of materials should be consistent or conform to a predetermined color range. If the material color becomes abnormal, such as darkening, lightening, or developing spots, it may indicate a chemical or physical change that could lead to adhesion or blockage. Using computer vision technology, the control device can analyze color histograms and identify abnormal patterns in color distribution. For example, if the color distribution of a certain area deviates significantly from the normal state, that area is marked and its level of attention is increased. Combining this with other parameters, such as adhesion level and flow rate, allows for a comprehensive assessment of the likelihood of blockage.

[0080] Texture information reflects the structure and pattern of the material surface. Changes in texture may indicate material accumulation or flow obstruction. Under normal flow conditions, the texture characteristics of materials exhibit a certain regularity and uniformity. Reduced texture uniformity may mean that materials are beginning to aggregate or clog. When materials begin to accumulate or clog in areas prone to blockage, the texture pattern on their surface will change, possibly manifesting as a denser, more irregular texture or the appearance of obvious blocky structures.

[0081] This embodiment captures material images using an onboard high-resolution camera and performs texture analysis on these images using computer vision technology. By calculating texture feature values, such as contrast, homogeneity, and entropy, the system can quantify the complexity and variability of the texture. When the detected texture features differ significantly from those during normal flow, the control device identifies a potential risk of blockage. For example, a decrease in texture entropy indicates that the material surface has become more uniform and lacks variation, which may be a precursor to blockage.

[0082] By integrating multiple monitoring data, such as adhesion level, flow rate, color information, and texture information, the control device can accurately determine whether easily clogged areas are blocked. This intelligent judgment mechanism greatly improves the timeliness of cleaning and reduces unnecessary cleaning operations caused by misjudgments, while ensuring the continuity and efficiency of the production process.

[0083] In some embodiments, step 103 may specifically include:

[0084] (1) Based on the degree of adhesion, flow rate, color information and texture information, determine the adhesion index B, flow rate index Q, color value C and texture entropy value T respectively.

[0085] The degree of adhesion is characterized by the bonding properties and distribution patterns of material particles. The distribution pattern can be represented by the number or proportion of adherent areas. The adhesion index B can be quantified by evaluating the bonding properties and distribution patterns of material particles in an image. The control device can employ image processing algorithms, such as edge detection and region growing, to identify contact points and adhesion areas between particles, thereby calculating the tightness of material adhesion and the size of the adhesion areas. A higher adhesion index B value indicates a more severe degree of adhesion between materials, potentially increasing the risk of clogging.

[0086] Edge detection aims to identify the location of object boundaries in an image. In material image analysis, edge detection can discover particle outlines and boundaries between particles. Commonly used edge detection algorithms include Sobel, Canny, and LaplaDIan. These algorithms determine edge locations by calculating the gradient of pixel intensity in the image. The output of edge detection is a binary image, where edge pixels are marked as white (value 1), and the remaining background pixels are marked as black (value 0).

[0087] In material images, contact points between particles often appear as abrupt changes in edges or sharp corners. Edge detection can identify these contact points and incorporate them as part of the bonding area. The results of edge detection can be used for subsequent image analysis, such as calculating the number and distribution of contact points to assess the adhesion tightness of the material.

[0088] Region growing is an image segmentation technique based on pixel similarity. It starts with one or more seed points and gradually merges adjacent pixels with similar attributes (such as color, brightness, texture, etc.) into the same region. In material image analysis, region growing can help identify and segment regions of adhered particles.

[0089] Region growing algorithms typically require defining a similarity criterion and a merge criterion. The similarity criterion determines which pixels can be considered similar and merged into the same region, while the merge criterion determines when to stop growing the region. The algorithm starts with a seed point, evaluates whether its neighboring pixels satisfy the similarity criterion, and if so, adds these pixels to the current region. This process is repeated until no more pixels satisfying the criteria can be added to any region.

[0090] Region growth yields a series of bonded regions, each representing a group of mutually bonded particles. By calculating the size, shape, and distribution of these regions, the material's bond strength can be assessed. For example, larger bonded regions may indicate higher bond strength, while smaller, more dispersed regions may indicate lower bond strength.

[0091] The following section details how to determine the adhesion index B based on the adhesion characteristics and distribution patterns of material particles:

[0092] 1) Preprocess the material images, including adjusting brightness and contrast and removing noise, to improve image quality.

[0093] 2) Apply edge detection algorithms (such as the Canny edge detector) to identify the boundaries of material particles in the material image, thereby determining possible adhesion points.

[0094] 3) Identify adhered regions from edge-detected images using a region growing algorithm. The specific algorithm steps are as follows: Select a set of seed points, which serve as the starting points for region growing. These seed points are typically located within the foreground or background region. Then, set growth rules that determine when pixels or regions can be merged. Common criteria include pixel grayscale similarity, texture consistency, and color similarity. Starting from a seed point, the algorithm checks its surrounding neighboring pixels and determines whether these pixels can be merged into the region where the seed point is located based on the growth criteria. If the criteria are met, the neighboring pixels are merged into the seed point's region, becoming new boundary points and added to the queue of pixels to be processed. Repeat the above region growing process until no new pixels can be merged, or the preset number of iterations is reached. The region growing algorithm can effectively merge physically adhered pixels into the same region, thereby identifying adhered material regions.

[0095] 4) Adhesion characteristics are evaluated by analyzing the number of contact points between particles. The more contact points, the greater the adhesion density and the more pronounced the adhesion characteristics. This allows for the evaluation of adhesion characteristics, yielding an assessment value. Specifically, an adhesion characteristic database can be used to obtain the corresponding assessment values. This database contains the correspondence between different adhesion characteristics and assessment values, and this correspondence can be established through expert evaluation and scoring of the adhesion characteristics.

[0096] 5) Calculate the number of pixels or the area ratio of the adhered region to obtain the evaluation value of the distribution pattern.

[0097] 6) The adhesion characteristic evaluation value and the distribution pattern evaluation value are weighted and summed to obtain the adhesion index B.

[0098] Flow rate describes the trajectory of material movement, and the flow rate index Q is determined by analyzing the material's trajectory across consecutive image frames. The control device uses object tracking technology to monitor the displacement of material particles across a series of image frames, thereby calculating the material's average velocity and flow rate.

[0099] The following details how object tracking technology is used to monitor and calculate the traffic index Q:

[0100] 1) Continuously capture image sequences of material flow using a high-speed camera to ensure a sufficiently high frame rate to capture the details of material movement.

[0101] 2) Preprocess each frame of the image, including noise reduction and contrast enhancement, in order to better identify and track material particles.

[0102] 3) Use object detection algorithms in each frame of the image to identify specific areas of the material flow.

[0103] 4) Apply object tracking techniques (such as Kalman filtering, Mean-Shift, or deep learning-based tracking algorithms) to monitor the displacement of material flow in consecutive image frames.

[0104] 5) Measure the displacement distance and time of the material flow between consecutive frames, and calculate the average velocity and acceleration of the material.

[0105] 6) Calculate the flow rate of the material based on its average velocity and the cross-sectional area of ​​the flow. The flow rate can be volumetric flow rate (e.g., cubic meters per second) or mass flow rate (e.g., kilograms per second).

[0106] 7) Convert the calculated flow rate value into a flow rate index Q that is convenient for subsequent calculations and processing. For example, a correlation between the flow rate value and the flow rate index Q can be preset, and then the corresponding flow rate index Q can be determined based on this correlation.

[0107] Color information describes the color characteristics of a material. Color values ​​(C) can quantify the color information of a material in various ways to facilitate computer analysis and processing. In computer vision and image processing, color is typically represented by color models, the most common of which include RGB (Red, Green, Blue), HSV (Hue, Saturation, Brightness), HSL (Hue, Saturation, Brightness), and CMYK (Cyan, Magenta, Yellow, and Black). To determine color values, a suitable color space must first be chosen. The RGB color space is the most basic color model, with each color channel typically ranging from 0 to 255 (for 8-bit color depth). The HSV color space is closer to how humans perceive color, where H represents hue (0 to 360 degrees), S represents saturation (0 to 100%), and V represents brightness (0 to 100%).

[0108] The following section details how to convert the color information of materials into quantifiable color values ​​C:

[0109] 1) Use a camera to capture images of the material surface in a controlled environment, ensuring consistent lighting conditions for accurate color comparison.

[0110] 2) Preprocess the image, including cropping (removing non-material parts from the image), scaling (adjusting the image size for easier processing), and conversion (such as converting from RGB color space to grayscale if color information is not the focus of the analysis).

[0111] 3) Choose a suitable color space for analysis. For most applications, HSV (Hue, Saturation, Brightness) or CIELAB color spaces better reflect human color perception than RGB.

[0112] 4) Within the selected color space, calculate the histogram for each color channel in the image. The histogram shows the distribution and quantity of each color in the image. The specific calculation method is as follows: First, determine the resolution of the histogram, that is, the number of bins (intervals in the histogram) for each channel, which will determine the precision of the histogram. Then, iterate through each pixel in the image, incrementing the count of the corresponding bin based on the pixel's value in each channel. Finally, construct the histogram using a mapping function from pixel value to bin index; for example, if the histogram for the H channel has 8 bins and the pixel's hue value is 30, then it is counted in the 4th bin (assuming 0 degrees corresponds to the first bin, and each bin covers 40 degrees).

[0113] 5) Extract features from the color histogram, such as:

[0114] Mean and standard deviation: describe the central location and dispersion of the color distribution.

[0115] Peak quantity and location: Indicates the dominant color in the image and its distribution.

[0116] Histogram entropy: describes the uniformity and complexity of color distribution.

[0117] 6) Convert the extracted histogram features into a single numerical value, namely the color value C. For example, you can take the weighted sum of the histogram features for each color channel, where the weights can be allocated according to the importance of the color channel.

[0118] Texture information represents the texture complexity of a material surface. The texture entropy value T can be calculated by analyzing the features of the material's texture information in an image. For example, the gray-level co-occurrence matrix (GLCM) can be used to extract features such as contrast and homogeneity of the texture, and then its entropy value can be calculated. A higher entropy value indicates a more random and complex texture; a lower entropy value indicates a more regular and simple texture. The texture entropy value can be calculated using the entropy formula in information theory.

[0119] The following section details how to calculate the texture entropy value T using the texture information of materials:

[0120] 1) Perform necessary preprocessing on the material image, such as converting it to grayscale, to facilitate texture analysis.

[0121] 2) Calculate the Gray-Level Co-occurrence Matrix (GLCM) based on the original image. The GLCM represents the gray-level relationship between pixel pairs in an image, including parameters such as angle, distance, and contrast. The specific calculation method is as follows: First, determine the distance (d) and direction (θ) used in calculating the GLCM. The distance refers to the distance between the center pixel and its neighboring pixels, while the direction defines the position of the neighboring pixels relative to the center pixel. Then, determine the number of gray levels used to calculate the GLCM. Gray levels are typically 8-bit, 16-bit, or other values, depending on the depth of the image. Then create another image... N × N The GLCM is a matrix where ∠p is the number of gray levels, and all matrix elements are initialized to 0. This matrix stores statistical information about the gray level pairs. Next, iterate through each pixel in the image. For each pixel P: calculate its gray value g; for each neighboring pixel P′ of P, calculate the gray value g′ of P′ based on the set direction and distance (d, θ). In the GLCM, find the matrix element corresponding to (g, g′) and increment its value by 1. Finally, normalize the GLCM by dividing each element in the GLCM by the total number of pixel pairs in the image, converting it into a joint probability distribution.

[0122] 3) Extract texture features from GLCM. Common features include:

[0123] Contrast: Indicates the difference in gray levels between different regions in an image.

[0124] Homogeneity: A statistical measure of how similar a pixel is to its neighbors in an image.

[0125] Energy (variance): Represents the degree of uniformity in image texture variation.

[0126] Correlation: Indicates the directionality of image texture.

[0127] 4) Calculate the entropy value of GLCM using the entropy formula from information theory. This entropy value is the texture entropy value T. The entropy formula is as follows:

[0128] ;

[0129] in, N is the number of gray levels in the GLCM, and p(i, j) is the probability value of the element in the i-th row and j-th column of the GLCM. This formula calculates the probability distribution entropy of the joint gray-level distribution based on the GLCM, which can be used as an indicator to measure the complexity and randomness of image texture, namely the texture entropy value T.

[0130] (2) Substitute the adhesion index B, flow index Q, color value C, and texture entropy value T into the clogging formula to calculate the clogging index DI. The mathematical expression of the clogging formula is as follows:

[0131] ;

[0132] Among them, B thr This represents the adhesion threshold, a predetermined threshold for the adhesion index. When the actual adhesion index is below this value, the material is considered unlikely to cause blockage; B max This represents the maximum possible value of the adhesion index B, used to normalize the effect of the adhesion index; Q nor This represents the theoretical normal flow rate of the easily clogged section under unobstructed conditions; K is the material flow rate attenuation coefficient, used to adjust the flow rate sensitivity, and can be adjusted according to actual conditions. C thr C represents the color threshold, indicating the critical point at which color changes. nor The normal color of the material, T thr T represents the texture entropy threshold, indicating the critical point at which the texture entropy value changes. nor Here, w1 represents the normal texture entropy value of the material, and w2, w3, and w4 are weighting coefficients. The values ​​of w1, w2, w3, and w4 all range from 0 to 1, and w1 + w2 + w3 + w4 = 1. By setting different weighting coefficients, the number of parameters in the formula and the influence of each parameter can be adjusted. For example, when w3 and w4 are both 0, this embodiment only considers the influence of adhesion degree and flow rate when determining whether there is blockage.

[0133] In the blocking formula:

[0134] It can measure the degree of adhesion relative to the adhesion threshold and the maximum possible value;

[0135] It can measure the attenuation of traffic flow relative to the theoretical normal traffic flow;

[0136] It can measure the change in color information relative to color thresholds and normal values;

[0137] It can measure the change in texture information relative to texture entropy and normal values.

[0138] This formula combines the material's adhesion, flow rate, color information, and texture information, and judges the blockage status by the relative changes of these four indicators.

[0139] (3) Determine whether the blockage-prone parts are in a blocked state based on the blockage index DI.

[0140] If the blockage index DI exceeds the preset warning value, it indicates that the adhesion is high and the flow rate is lower than normal. The control device will determine that the blockage-prone part may be blocked.

[0141] For example, suppose the warning value for the congestion index DI is 0.6.

[0142] The adhesion index B = 0.4, B thr =0.1, B max =1, flow index Q=0.6, Q nor =1, K=0.5, color value C=0.15, C thr =0.5, C nor =0.3, texture entropy T=0.25, T thr =0.1, T nor =0.3.

[0143] w1=0.4, w2=0.4, w3=0.1, w4=0.1.

[0144] Substituting all the above parameters into the blockage formula, we can obtain a blockage index DI of 0.69, which is greater than the warning value of 0.6. Therefore, we can determine that the easily blocked part is in a blocked state.

[0145] Step 104: If the blockage is in progress, determine the target cleaning plan based on the material characteristics and flow rate.

[0146] Specifically, when the control device detects that the easily clogged part of the feed trough is stuck, it will determine the target cleaning plan based on the material characteristics such as the degree of adhesion, flow rate, color information and texture information of the material.

[0147] For example, the control unit uses computer vision technology to analyze images captured by cameras to assess the tightness of material adhesion and the degree of flow obstruction. If the adhesion is high, the control unit will instruct the robot to use moderate cleaning force to ensure effective separation of the adhered material without damaging the feed trough. Simultaneously, depending on the degree of flow reduction, the control unit will adjust the robot's movement speed, slowing it down to allow more time for the cleaner to work on the adhered area, or appropriately increasing the speed to improve cleaning efficiency if the flow is still acceptable.

[0148] The control unit optimizes the robot's movement trajectory based on the distribution and degree of adhesion and blockage, ensuring that the robot can cover all critical areas and prioritize cleaning the most severe blockages. The operating frequency is also adjusted according to the adhesion condition; for stubborn adhesions, the operating frequency is increased to improve cleaning effectiveness; for looser materials, the frequency is reduced to avoid over-cleaning. Through these comprehensive adjustments, the control unit can formulate a precise target cleaning plan, quickly and effectively clearing blockages, restoring normal material flow, and minimizing disruption to the production process.

[0149] The control device can also identify potential blockages by analyzing changes in the color of the material surface in real time. For example, if a significant deviation is detected in the material color from a reference color under normal conditions, this may indicate that the material is beginning to adhere. In this case, the cleaning device can be instructed to increase the frequency or intensity of cleaning the area, using specific cleaning tools or detergents to remove the adhered material indicated by the color change.

[0150] The control unit can also assess the flow state and potential clogging risk of materials by analyzing the texture features of the material surface. When the texture of the material surface deviates from expectations, it may mean that material is beginning to accumulate or flow is obstructed. In this case, the cleaning device can adjust its cleaning path, focusing on areas with significant texture changes, and may need to change its cleaning mode, such as using high-frequency or large-amplitude cleaning motions to break up adhering materials, or adjusting the angle and pressure of the cleaning device to more effectively remove blockages and promote material flow. Through this intelligent adjustment based on texture information, the cleaning robot can more accurately cope with various cleaning challenges, ensuring the smooth operation of the production process.

[0151] If it is determined that there is no blockage, return to step 101 and continue monitoring.

[0152] Step 105: Generate the first cleaning control command based on the target cleaning plan.

[0153] The first cleaning control command is used to control the cleaning device to perform cleaning operations in order to achieve the target cleaning plan.

[0154] Specifically, the control device determines the target cleaning plan for the jamming situation, and it will generate a series of detailed first cleaning control commands based on the plan. These commands include specific operational requirements for the cleaning device, such as the set value of the cleaning force, the adjustment parameters of the moving speed, the optimization value of the action frequency, and the precise path planning of the moving trajectory.

[0155] The control device first converts the parameters in the target cleaning plan into executable control signals. These signals guide the drive motor of the cleaning device to move the sweeper with appropriate force and speed. For example, if it is necessary to increase the cleaning force, the control device will send a command to increase the output power of the motor; if it is necessary to adjust the moving speed, the control device will adjust the motor speed to change the moving rate of the sweeper.

[0156] Simultaneously, the control device sets the operating frequency of the sweeping device to ensure that the sweeper swings or rotates at the most efficient frequency to achieve the best cleaning effect. In addition, the control device generates navigation commands based on the predetermined movement trajectory, guiding the sweeping device to the easily clogged areas and bypassing obstacles when necessary.

[0157] These initial cleaning control commands are sent to the cleaning device in real time via a communication interface to ensure precise execution of the cleaning operation and achieve the target cleaning plan. The control device also continuously monitors the cleaning process and adjusts the commands based on actual feedback to optimize the cleaning effect and ensure the smooth completion of the cleaning task. In this way, the control device ensures that the cleaning device can perform cleaning operations efficiently and accurately, effectively clearing blockages and keeping the feed chute unobstructed.

[0158] Step 106: Send the first cleaning control command to the cleaning device.

[0159] Specifically, after generating the first cleaning control command, the control device transmits these commands to the cleaning device via a communication interface. The communication interface can be a wired connection, such as Ethernet, serial communication, or USB, or a wireless connection, such as Wi-Fi, Bluetooth, or other wireless protocols. These control commands contain parameters such as cleaning intensity, movement speed, motion frequency, and movement trajectory, and they are encoded into a format that the cleaning device can understand and execute.

[0160] During transmission, the control device may employ one or more communication protocols to ensure accurate data transmission and reception. For example, it may use the TCP / IP protocol to establish a reliable connection or the UDP protocol to send real-time data. Control commands may be encrypted before transmission to ensure operational security and prevent unauthorized access or interference.

[0161] After receiving control commands, the cleaning device's built-in microcontroller or computer system parses these commands and converts them into corresponding actions. The drive motor adjusts its speed and torque according to the commands to execute the sweeper's movement and cleaning actions. Simultaneously, the position sensors and status indicators on the cleaning device feed back the execution status to the control unit, forming a closed-loop control system to ensure that the cleaning operation is carried out accurately and without error according to the predetermined target cleaning plan.

[0162] Through this efficient communication and command transmission mechanism, the control device can control the cleaning device in real time and accurately, ensuring the smooth completion of the cleaning task and solving the problem of blockage in the feeding trough in a timely and effective manner.

[0163] This embodiment acquires material images of easily clogged areas and analyzes them using computer vision technology, achieving accurate identification of material adhesion, flow rate, color information, and texture information. This intelligent identification system can monitor the material status in real time and promptly detect potential blockage problems. Combining this material information, the system can accurately determine whether there is a risk of blockage at easily clogged areas, thereby taking preventive measures. When a blockage is detected, the system automatically formulates a suitable cleaning plan and generates corresponding control commands. These commands guide the cleaning device to perform effective cleaning operations, ensuring that blockages are cleared in a timely manner and guaranteeing smooth production processes. The entire process is automated and intelligent, significantly improving cleaning efficiency and accuracy, reducing manual intervention, lowering production costs, eliminating the need for air cannons or vibration devices for unblocking, reducing damage to production equipment during cleaning, achieving intelligent maintenance and management of the feed trough, and improving the overall system's operational safety and reliability.

[0164] In some embodiments, the cleaning scheme can be executed at four different times: First, when the feed chute becomes clogged, cleaning can be started immediately to remove the blockage and restore normal operation. Second, regular cleaning can be performed, such as cleaning every three days or once a week, to keep the feed chute unobstructed for a long time. Third, pre-cleaning can be performed when the material is found to be highly adhesive to prevent blockage. Fourth, maintenance cleaning can be performed after the material transportation is completed to reduce material residue in easily clogged areas.

[0165] In some embodiments, the cleaning system further includes a lidar, a humidity sensor, and a flow sensor. The lidar is used to detect the three-dimensional shape information of the material at the easily clogged parts, and the humidity sensor and flow sensor are installed at the easily clogged parts of the feed chute.

[0166] To further demonstrate the advantages of the method in this embodiment, we will now combine... Figure 5 To explain this method in detail, the following steps are included:

[0167] Step 201: Obtain material images of easily clogged areas and acquire three-dimensional shape information of the material using lidar.

[0168] Among them, lidar calculates distance by emitting laser pulses and measuring the time it takes for the light reflected back from the material, thereby obtaining precise location and shape information of the material at easily clogged areas.

[0169] LiDAR can capture the three-dimensional shape of an object, providing more comprehensive spatial information. This three-dimensional shape information can characterize the specific shape and size of the material, thus determining whether there are foreign objects or large pieces of material in a potentially clogged area. For example, if the detected shape differs significantly from the material, it can be determined that there may be foreign objects in the potentially clogged area; or if the detected volume is large, it can be determined that there may be large pieces of material in the potentially clogged area.

[0170] Step 202: Based on computer and vision technology, identify material images to obtain the adhesion degree, color information, and texture information of the material in the easily clogged areas, as well as the flow rate of the material inlet and outlet. Combine the adhesion degree, color information, texture information, and three-dimensional shape information to obtain material feature information.

[0171] By using cameras to simultaneously identify adhesion levels, flow rates, color information, and texture information, combined with precise 3D shape information acquired by LiDAR, the control device can achieve comprehensive monitoring and in-depth analysis of material conditions. This multi-dimensional data fusion method greatly improves the accuracy of identifying material flow and potential blockage problems, enabling the system to promptly detect and respond to complex situations such as adhesion, foreign object intrusion, and large material accumulation. Intelligent analysis algorithms can adjust cleaning strategies based on real-time data, optimizing cleaning paths and operating modes, thereby effectively preventing and resolving blockage problems, ensuring continuous operation and production efficiency of the production line, while also reducing the need for manual inspection and intervention, and improving operational safety and intelligence.

[0172] Step 203: Based on the material characteristics and flow rate, determine whether the blockage-prone parts are in a blocked state.

[0173] First, image processing techniques are used to identify and quantify adhesion areas on the material surface. If the degree of adhesion exceeds a preset threshold, it may indicate a risk of clogging. Second, real-time tracking of material flow rate and velocity using cameras reveals significant drops or instabilities in flow rate, also suggesting potential clogging issues. Abnormal changes in color information, such as shifts in color histograms, may reveal alterations in the chemical or physical properties of the material due to humidity or compression, leading to clogging. Texture analysis reveals changes in the microstructure of the material surface; abnormal variations in texture uniformity and complexity may indicate material accumulation. Three-dimensional shape information provided by LiDAR helps identify areas and morphologies of material accumulation; irregular three-dimensional structures or abnormal volume growth are strong evidence of clogging.

[0174] By comprehensively evaluating the above data, we can accurately determine the condition of easily clogged areas and take timely cleaning measures to prevent or resolve blockages, ensuring smooth and efficient production processes.

[0175] Specifically, the following steps are used to integrate the material's adhesion, flow rate, color information, texture information, and three-dimensional shape information to comprehensively determine whether the easily clogged parts are blocked:

[0176] (1) Data standardization: Standardize the collected data (adhesion degree, flow rate, color information, texture information and three-dimensional shape information) to ensure that they are on the same order of magnitude, which facilitates comparison and fusion. For example, all data can be scaled to the range of 0 to 1.

[0177] Specifically, the min-max scaling method can be used to standardize these data. For a given data point, subtract its minimum value and divide by the difference between its maximum and minimum values.

[0178] For example, the standardization formula is: X norm =( X - X min ) / ( X max - X min );

[0179] in, X It is the raw data. X min It is the minimum value of this feature. X max It is the maximum value of this feature. X norm It is standardized data.

[0180] Through this process, the data of all features will be converted to the same numerical range, which facilitates subsequent fusion processing.

[0181] (2) Feature weighting: Assign weights to each feature based on experience or expert knowledge to reflect their importance in blockage determination. For example, the degree of adhesion and flow conditions may be more critical than color and texture information.

[0182] (3) Calculation of comprehensive score: Calculate the weighted score of each feature according to the assigned weight. For example, you can multiply the current value of each feature by its weight, and then add up the weighted scores of all features to get a comprehensive score.

[0183] (4) Threshold setting: Set a comprehensive scoring threshold to determine whether a blockage has occurred. The threshold can be determined based on historical data and actual operating conditions.

[0184] (5) Definition of logical rules: Define a set of logical rules to determine the blockage status based on the comprehensive score and the status of each feature. For example, if the comprehensive score exceeds the threshold, or if the adhesion degree and flow status both indicate blockage, then it is judged to be in a blockage state.

[0185] (6) Real-time monitoring and response: Monitor the data of each feature in real time and calculate the comprehensive score periodically. Assess whether the element is in a blocked state based on the comprehensive score.

[0186] If the system is determined to be in a blocked state, proceed to step 204; otherwise, proceed to steps 207 and 209.

[0187] Step 204: Determine the target cleaning plan based on material characteristics and flow rate.

[0188] This step is the same as step 104, and will not be repeated here.

[0189] Step 205: Generate a first cleaning control command based on the target cleaning plan. The first cleaning control command is used to control the cleaning device to perform cleaning operations to achieve the target cleaning plan.

[0190] This step is the same as step 105, and will not be repeated here.

[0191] Step 206: Send the first cleaning control command to the cleaning device. Then proceed to step 211.

[0192] This step is the same as step 106, and will not be repeated here.

[0193] Step 207: Obtain material moisture information for easily clogged areas. Proceed to step 208.

[0194] The material humidity information is obtained by a humidity sensor.

[0195] The control device acquires the material's humidity information by communicating with humidity sensors installed at locations prone to clogging. These humidity sensors are specifically designed to monitor and measure the material's humidity level in real time and transmit the collected data back to the control device. Upon receiving this data, the control device can analyze the material's humidity status in real time to determine if there is a risk of clogging due to excessive humidity.

[0196] Step 208: When the material humidity information is greater than the set humidity threshold, a second cleaning control command is sent to the cleaning device. Then proceed to step 211.

[0197] The second cleaning control command is used to control the cleaning device to perform a preset cleaning operation.

[0198] When the control device detects that the material moisture content at a clog-prone area exceeds a preset moisture threshold, it immediately sends a second cleaning control command to the cleaning device. This command is based on a pre-designed response strategy for high humidity conditions and aims to initiate and guide the cleaning device to perform a series of preset cleaning operations. These operations may include increasing the cleaning frequency, adjusting the cleaning intensity, or adopting a specific cleaning mode to more effectively handle moist materials and reduce the risk of clogging. Upon receiving the command, the cleaning device automatically executes the corresponding cleaning task to ensure that the material moisture content is controlled, thereby maintaining the unobstructed flow of the feed chute and the stable operation of the production process.

[0199] Step 209: Obtain material flow information for easily clogged areas. Proceed to Step 210.

[0200] The material flow rate information is obtained by a flow sensor.

[0201] The control device connects to flow sensors installed in the feed chute system to acquire real-time material flow information at potentially clogged sections. The flow sensors monitor and measure the amount of material passing through a specific cross-section per unit time, providing crucial flow data to the control device. If the control device detects that the material flow rate is below a set threshold, this may indicate that material flow is obstructed or there is a risk of blockage.

[0202] Step 210: When the material flow rate is less than the set flow rate threshold, a third cleaning control command is sent to the cleaning device. Then proceed to step 211.

[0203] The third cleaning control command is used to control the cleaning device to perform preset cleaning operations.

[0204] If the material flow rate is less than the set flow threshold, the control device will send a third cleaning control command to the cleaning device. This command is based on a predefined cleaning strategy and operating procedure, and is designed to guide the cleaning device to perform targeted cleaning operations. These operations may include adjusting the movement trajectory of the cleaning device, the cleaning intensity, or activating a specific cleaning mode to effectively remove potential blockages and restore normal material flow.

[0205] Upon receiving a third cleaning control command, the cleaning device automatically executes the corresponding cleaning tasks, such as activating the robotic arm and brushes, adjusting the scraper position, or activating the vacuum system, to ensure that easily clogged areas are cleaned promptly, preventing further blockages and thus ensuring the continuity and efficiency of the entire production process. Through this automated response mechanism, the system can quickly identify and resolve potential blockages, reducing the risk of production interruptions.

[0206] Step 211: After receiving the cleaning completion information sent by the cleaning device, acquire the image of the cleaned material captured by the camera.

[0207] Upon receiving the cleaning completion notification from the cleaning device, the control device sends a command to the camera to immediately capture images of the cleaned material. The camera then captures the images and sends them back to the control device. These images are crucial for evaluating the cleaning effectiveness; they are used to determine if the cleaning has met the expected standards, ensuring that easily clogged areas have been thoroughly cleaned, thus preventing future blockages. By analyzing these images, the control device can check the distribution of material, confirm the reduction in adhesion, and the removal of clogging substances, thereby verifying the cleaning device's efficiency and cleaning quality.

[0208] Step 212: Determine whether the cleaning result of the cleaning device is qualified based on the image of the cleaned material.

[0209] After acquiring images of the cleaned material, the control device will use computer vision technology and image processing algorithms to perform in-depth analysis of the images. It will identify and compare the distribution, adhesion, and residue of the material in the images before and after cleaning, and determine whether the cleaning results meet the requirements based on set acceptance criteria (such as uniform material distribution, no obvious adhesion areas, and no significant residue).

[0210] If the image analysis results show that the state of the cleaned material meets the preset cleaning standard, then the cleaning device is considered to have completed the cleaning task successfully.

[0211] Conversely, if the analysis results show that the cleaning is incomplete or has not achieved the expected results, the control device will mark the cleaning result as unqualified, which may trigger further cleaning procedures or prompt manual intervention, and proceed to step 213. This process ensures the quality and efficiency of the cleaning operation and guarantees the smooth operation of the production process.

[0212] Step 213: Adjust the target cleaning plan based on the image of the cleaned material.

[0213] After receiving and analyzing images of the cleaned material, if the control device finds that the cleaning results do not meet the predetermined standards, it will adjust the target cleaning plan based on the material distribution, adhesion, and specific characteristics of residual blockages shown in the images. This may include resetting cleaning intensity parameters, modifying the cleaning path to cover incompletely cleaned areas, increasing cleaning frequency or duration, or even changing the type or configuration of the cleaning equipment. The adjusted plan aims to more effectively remove residues, ensure unobstructed flow in the feed chute, optimize energy and resource usage, and improve the overall efficiency and effectiveness of the cleaning operation. The control device translates these adjustments into specific operating instructions, guiding the cleaning equipment to perform necessary secondary cleaning or other supplementary cleaning actions.

[0214] In some embodiments, after step 103, the following steps are further included:

[0215] (1) Based on the current material humidity and material flow information of the easily blocked parts, determine the estimated time of blockage when the easily blocked parts will become blocked.

[0216] The material humidity information reflects the moisture content of the material. High humidity may cause the material to adhere to the tank or stick together, thus increasing the risk of blockage. The material flow rate information shows the amount of material passing through the blockage-prone area per unit time. Abnormal changes in flow rate, such as a significant decrease, may indicate that the material flow is obstructed.

[0217] The control unit combines these two key parameters, analyzes historical data and material characteristics, and uses machine learning models, such as time series analysis or regression models, to predict the occurrence of blockages. For example, if the current humidity is close to or exceeds the levels seen during historical blockage events, and the flow rate shows a downward trend, the control unit will calculate a possible blockage time window. This time window, based on a probabilistic model, provides the likelihood and urgency of the blockage.

[0218] In this way, the control device can identify potential clogging risks in advance and take preventative measures before the estimated clogging time. This allows the cleaning device to intervene promptly and perform cleaning tasks, effectively avoiding or mitigating clogging events and ensuring the continuity and efficiency of the production process. This intelligent prediction and prevention strategy significantly improves the initiative and responsiveness of feeder management.

[0219] In this embodiment, the process of using a machine learning model to predict material blockage time can be divided into the following key steps:

[0220] Data Collection and Analysis: Collect historical data, including parameters such as material humidity, flow rate, particle size distribution, and conveying speed. Analyze the data characteristics of the material under normal flow and clogging conditions to identify key factors that may affect clogging.

[0221] Feature engineering involves extracting useful features from raw data, such as statistical characteristics of humidity and flow (mean, variance, peak value, etc.) and time series characteristics (trend, periodicity, etc.). External factors, such as ambient temperature and pressure, as well as equipment operating parameters, may also need to be considered.

[0222] Model selection and training: Choose a suitable machine learning model, such as decision tree, random forest, support vector machine, neural network, etc. Train the model using historical data and optimize model parameters through methods such as cross-validation to ensure that the model has good generalization ability.

[0223] Model Evaluation: Evaluate the model's predictive performance using a test set, measuring its predictive effectiveness using metrics such as accuracy, recall, and F1 score. If possible, real-world congestion events can be used to validate the model's predictive accuracy.

[0224] Prediction and Implementation: Deploy the trained model to the production environment to monitor the flow of materials in real time. Predict the timing of blockages based on the model's output and take timely preventative measures when potential blockage risks are predicted.

[0225] Feedback and Iteration: Collect model prediction results and actual congestion event data, and analyze the model's prediction error. Adjust model features and parameters based on feedback, and continuously iterate to optimize the model, improving the accuracy and reliability of predictions.

[0226] (2) Determine the cleaning task execution time based on the estimated blockage time.

[0227] Specifically, the control device will determine a suitable time to execute the cleaning task. This time point will be selected some time before the blockage occurs to ensure that the cleaning operation can be carried out before or in the early stages of blockage, thereby effectively preventing or mitigating the impact of the blockage. The determination of the execution time will also take into account the complexity of the cleaning operation and the preparation time of the cleaning device, ensuring that the cleaning device has sufficient time to start and reach the designated location.

[0228] In some embodiments, this step includes: determining an off-peak electricity period prior to the estimated congestion time; and determining the cleaning task execution time within the off-peak electricity period.

[0229] Assuming that a machine learning model predicts the feed chute blockage will occur at 10:00 AM the following day, historical data analysis shows that off-peak electricity hours are typically between 1:00 AM and 7:00 AM, when prices are about 50% lower than usual.

[0230] In this case, the duration of the cleaning operation needs to be considered. Assuming a thorough cleaning operation takes 1 hour to complete, it is desirable to start cleaning before 7:00 AM to complete most of the work during off-peak electricity hours, while ensuring sufficient time to complete the cleaning and resume production before the estimated blockage time.

[0231] Therefore, you can choose to start the cleaning task at 5:00 AM. This way, the cleaning will be completed by 6:00 AM, which is 4 hours before the estimated blockage time, enough to handle any emergencies or additional cleaning needs. At the same time, this time falls within off-peak electricity hours, saving on electricity costs.

[0232] In this way, cleaning can be completed before blockages occur, ensuring smooth production processes, and cleaning tasks can be performed during periods of lower electricity costs, thereby reducing overall operating costs.

[0233] (3) When the cleaning task execution time is reached, the first cleaning control command is sent to the cleaning device.

[0234] Specifically, when the predetermined cleaning task execution time arrives, the control device generates specific first cleaning control commands based on the previously analyzed and formulated cleaning strategy. These commands include operational parameters such as the sequence of actions the cleaning device needs to perform, cleaning intensity, movement path, and speed. Subsequently, the control device transmits these commands to the cleaning device via a wired or wireless communication network. Upon receiving the commands, the cleaning device starts its drive motor and executes cleaning operations according to the parameters specified in the commands, such as adjusting the movement of the robotic arm, the rotation speed of the brushes, activating the vacuum system, or deploying other cleaning tools, moving along the planned path to efficiently and accurately remove material from easily clogged areas and ensure unobstructed flow in the feed chute. Throughout the cleaning process, the control device continuously monitors the status of the cleaning device and the cleaning effect, adjusting commands as necessary to optimize cleaning performance.

[0235] The control device in this embodiment is an electronic device. The electronic device of this invention embodiment is described below from a hardware processing perspective. Please refer to [link / reference]. Figure 6 This is a schematic diagram of the physical device structure of an electronic device in an embodiment of this application.

[0236] It should be noted that, Figure 6 The structure of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of the present invention.

[0237] like Figure 6As shown, the electronic device includes a Central Processing Unit (CPU) 401, which can perform various appropriate actions and processes based on a program stored in Read-Only Memory (ROM) 402 or a program loaded from storage portion 408 into Random Access Memory (RAM) 403, such as performing the methods described in the above embodiments. The RAM 403 also stores various programs and data required for system operation. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An Input / Output (I / O) interface 405 is also connected to the bus 404.

[0238] The following components are connected to I / O interface 405: input section 406 including audio input devices, push-button switches, etc.; output section 407 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 408 including a hard disk, etc.; and communication section 409 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0239] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs the various functions defined in the present invention.

[0240] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0241] 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 the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0242] Specifically, the electronic device of this embodiment includes a processor and a memory. The memory is coupled to one or more processors and is used to store computer program code. The computer program code includes computer instructions. One or more processors call the computer instructions to cause the electronic device to perform the method provided in the above embodiment.

[0243] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The storage medium carries one or more computer programs that, when executed by a processor of the electronic device, cause the electronic device to implement the methods provided in the above embodiments.

[0244] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0245] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0246] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A self-identifying intelligent cleaning method for enclosed spaces, characterized in that, A control device for a cleaning system, the cleaning system further comprising a cleaning device and a camera, the cleaning device being installed at a clog-prone section of the feeding trough, the camera being positioned to capture images of the clog-prone section and the material inlet and outlet of the feeding trough, the cleaning device and the camera being communicatively connected to the control device; the method comprising: The system acquires material images of the easily clogged areas and the material inlets and outlets, the material images being captured by the camera. The material image is identified using computer and vision technology to obtain material characteristic information of the easily clogged parts and the flow rate of the material inlet and outlet. The material characteristic information includes at least one of adhesion degree, color information and texture information. Determining whether the easily clogged part is in a blocked state based on the material characteristic information and the flow rate includes: determining the adhesion index B, flow rate index Q, color value C, and texture entropy value T based on the adhesion degree, the flow rate, the color information, and the texture information, respectively. Substituting the adhesion index B, the flow index Q, the color value C, and the texture entropy value T into the clogging formula, the clogging index DI is calculated. The mathematical expression of the clogging formula is as follows: Among them, B thr B represents the adhesion threshold. max Q represents the maximum possible value of the adhesion index B. nor This represents the theoretical normal flow rate of the easily clogged section under unblocked conditions, where K is the material flow rate attenuation coefficient, and C... thr C represents the color threshold, indicating the critical point at which color changes. nor The normal color of the material, T thr T represents the texture entropy threshold, indicating the critical point at which the texture entropy value changes. nor The normal texture entropy value of the material is represented by w1, w2, w3, and w4, which are weighting coefficients. The blockage index DI is used to determine whether the easily blocked part is in a blocked state. If so, then the target cleaning plan is determined based on the material characteristic information and the flow rate. A first cleaning control instruction is generated according to the target cleaning plan. The first cleaning control instruction is used to control the cleaning device to perform cleaning operations in order to achieve the target cleaning plan. Send the first cleaning control command to the cleaning device; If it is not in a blocked state, the estimated time of blockage at the blocked part is determined based on the current material humidity and material flow information of the blocked part. Determine the off-peak electricity period before the estimated blockage time; determine the cleaning task execution time within the off-peak electricity period; When the cleaning task execution time arrives, a cleaning instruction is sent to the cleaning device.

2. The method according to claim 1, characterized in that, The cleaning system also includes a lidar system, and the material characteristic information includes the three-dimensional shape information of the material. Before the step of determining whether the easily clogged part is in a clogged state based on the material characteristic information and the flow rate, the system further includes: The lidar acquires the three-dimensional shape information of the material at the easily clogged location, and the three-dimensional shape information is used to determine whether there are foreign objects or large pieces of material at the easily clogged location.

3. The method according to claim 1, characterized in that, The cleaning system also includes a humidity sensor, which is installed at the clog-prone area; after the step of determining whether the clog-prone area is blocked based on the material characteristic information and the flow rate, the system further includes: If the easily clogged part is not in a clogged state, the material humidity information of the easily clogged part is obtained, and the material humidity information is collected by the humidity sensor; When the material humidity information is greater than the set humidity threshold, a second cleaning control command is sent to the cleaning device. The second cleaning control command is used to control the cleaning device to perform a preset cleaning operation.

4. The method according to claim 1, characterized in that, The cleaning system also includes a flow sensor, which is installed at the clogging-prone area; after the step of determining whether the clogging-prone area is blocked based on the material characteristic information and the flow rate, the system further includes: If the easily clogged part is not in a clogged state, the material flow information of the easily clogged part is obtained, and the material flow information is collected by the flow sensor; When the material flow rate information is less than the set flow rate threshold, a third cleaning control command is sent to the cleaning device. The third cleaning control command is used to control the cleaning device to perform a preset cleaning operation.

5. The method according to any one of claims 1-4, characterized in that, After the step of sending the first cleaning control command to the cleaning device, the method further includes: After receiving the cleaning completion information sent by the cleaning device, the image of the cleaned material captured by the camera is obtained; Determine whether the cleaning result of the cleaning device is qualified based on the image of the cleaned material. If not, the target cleaning plan is adjusted based on the image of the cleaned material.

6. A control device, characterized in that, include: Memory and one or more processors; The memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code including computer instructions, the one or more processors invoking the computer instructions to cause the control device to perform the method as described in any one of claims 1-5.

7. A computer-readable storage medium storing computer instructions, characterized in that, When the computer instructions are executed on the control device, the control device performs the method as described in any one of claims 1-5.

8. A computer program product, characterized in that, When the computer program product is run on the control device, the control device performs the method as described in any one of claims 1-5.

Images