A blanking hole clogging detection system based on infrared thermal imaging

Abstract

The application provides a blanking hole clogging detection system based on infrared thermal imaging. The blanking hole is adjacent to the end of a material conveying belt. The thermal imaging device of the infrared image acquisition unit comprises a long-wave infrared detector arranged at the blanking hole through a mounting bracket. The long-wave infrared detector can obtain a continuous infrared image sequence in a high-concentration dust environment. The field angle of the long-wave infrared detector is adjusted to cover the entire blanking hole area, so that an abnormal heat area formed when the blanking hole is clogged enters the imaging range completely. When the detection system is working, the intelligent computing unit analyzes the infrared thermal image sequence, analyzes the spatial distribution characteristics and time evolution characteristics of the temperature field, and determines whether the blanking hole is clogged by using a multi-feature fusion diagnosis model. The application can accurately and continuously collect and continuously monitor the infrared light data at the blanking hole in a high-dust-concentration environment and under the condition that the infrared light characteristics of different materials of the conveying belt are seriously different, and determine whether the blanking hole is clogged abnormally.

Description

Technical Field

[0001] This invention relates to the field of infrared light intensity detection technology in equipment condition assessment, and in particular to a material blockage detection system for material discharge port based on infrared thermal imaging, and also relates to the status monitoring and safety control of industrial material conveying processes. Background Technology

[0002] In bulk material conveying systems, the discharge port is a critical location prone to blockage. When material accumulates at the discharge port due to changes in particle size, increased moisture content, foreign matter contamination, or equipment structural issues, it not only reduces conveying efficiency but can also cause belt misalignment, tearing, or even equipment damage, posing significant safety hazards. Existing blockage detection technologies mainly include the following three categories:

[0003] The first category is mechanical contact detection methods, such as levers, baffles, and limit switches. These methods have simple structures, but their operation depends on the direct action of materials on the mechanical structure. They are prone to failure in high-dust and high-impact environments, and their response has a significant lag.

[0004] The second type is single-point temperature or pressure detection, such as installing thermocouples or temperature sensors on the wall of the material inlet. When the detected temperature exceeds a set threshold, it is determined that the material is blocked. However, this method can only obtain temperature information at a local point and cannot reflect the overall thermal field distribution. When the ambient temperature fluctuates or the initial temperature of the material changes, it is very easy to produce false alarms or missed alarms.

[0005] The third category is visual or laser detection methods, which use ordinary visible light cameras or laser sensors to detect the material accumulation pattern. However, in actual industrial settings, factors such as dust, steam, and changes in lighting conditions severely affect image quality, making this type of solution unstable.

[0006] Compared to the methods mentioned above, infrared thermal imaging technology can directly acquire temperature distribution information of the target area without contacting the material, making it theoretically very suitable for material blockage detection. However, most existing infrared material blockage detection solutions are still at the level of "maximum temperature alarm" or "average temperature threshold judgment," failing to fully explore the information contained in the spatial structure and temporal evolution of infrared thermal images, resulting in a high false alarm rate under complex working conditions.

[0007] Therefore, there is an urgent need for a blockage detection system that can fully utilize the spatiotemporal features of infrared thermal imaging and combine them with intelligent analysis methods to achieve earlier, more accurate, and more stable blockage identification. Summary of the Invention

[0008] This invention proposes a material blockage detection system for the material discharge port based on infrared thermal imaging. It can accurately and continuously collect and monitor the infrared light intensity distribution data and change characteristics data at the material discharge port in environments with high dust concentration and significant differences in infrared light characteristics of different materials on the conveyor belt. Furthermore, it can use the infrared light monitoring data to determine whether there is a material blockage abnormality at the material discharge port.

[0009] The present invention adopts the following technical solution.

[0010] A material blockage detection system based on infrared thermal imaging can perform material blockage detection by measuring the movement state of loose materials at the material inlet using infrared light in a high-concentration dust environment. The detection system includes an infrared image acquisition unit and an intelligent computing unit.

[0011] The material discharge port is adjacent to the end of the material conveyor belt. The thermal imaging device of the infrared image acquisition unit includes a long-wave infrared detector installed at the material discharge port via a mounting bracket. The working wavelength range of the long-wave infrared detector is set to 8-14μm to acquire continuous infrared image sequences in high-concentration dust environments. This is to adapt to the environment of large dust concentrations and difficult visible light imaging in industrial sites where bulk material is transported. The field of view of the long-wave infrared detector is adjusted to cover the entire material discharge port area and its nearby key parts, so that the abnormal heat zone formed when the material discharge port is blocked can be fully included in the imaging range.

[0012] When the detection system is working, the long-wave infrared detector continuously outputs an infrared image sequence corresponding to the two-dimensional temperature field distribution at the material discharge port to the intelligent computing unit, denoted as I(x,y,t), where x and y represent the pixel coordinates in the image and t represents the time series index. First, the two-dimensional temperature field distribution information at the material discharge port is measured by the long-wave infrared detector to generate an infrared thermal image sequence. The intelligent computing unit analyzes the temperature gradient, hot zone area, temperature mean or fluctuation of the infrared thermal image sequence to obtain the temperature distribution at the material discharge port, the trend of continuous heat accumulation in local areas, and the trend of temperature field changes over time. It also performs joint analysis on the spatial distribution characteristics and temporal evolution characteristics of the temperature field, and adopts a multi-feature fusion diagnostic model that combines rule judgment and intelligent algorithms to accurately identify the blockage state at the material discharge port. Finally, the early warning execution unit outputs an alarm signal or linkage control command.

[0013] The long-wave infrared detector is installed on one side of the outer casing of the thermal imaging equipment, and a dust cover is provided at the detection end of the long-wave infrared detector.

[0014] The intelligent computing unit includes an embedded platform with AI computing capabilities, used to process the calculation process of the multi-feature fusion diagnostic model. The intelligent computing unit is installed on the other side of the device casing. The external interfaces of the intelligent computing unit include a power input and an RJ45 interface.

[0015] The thermal imaging equipment is mechanically fixed above or to the side of the material inlet with a metal bracket. The outer shell of the thermal imaging equipment is a dustproof, shockproof, and high-temperature resistant industrial protective shell.

[0016] If the infrared radiation and reflection characteristics of different materials or discharge port structures vary significantly when the material conveyor belt is conveying different bulk materials, and this directly and seriously affects the accuracy and reliability of the infrared light intensity detection of the discharge port blockage detection system, then the multi-feature fusion diagnostic model of the detection system, based on the extraction and fusion of spatiotemporal thermal imaging features, further combines the infrared characteristics of different bulk materials with the reflection characteristics of the discharge port structure to achieve precise adaptation and optimization for various industrial scenarios, thereby improving the system's practicality and robustness; the measurement, calibration, and sampling data calculation of the infrared light intensity distribution and changes at the discharge port are performed, specifically including:

[0017] Method A, Material Infrared Characteristic Classification and Feature Calibration, specifically addresses the significant differences in infrared radiation and reflection characteristics among different materials during bulk material transport, directly impacting the accuracy and reliability of thermal imaging detection. For instance, highly reflective materials (such as metal ores and dry sand) readily reflect ambient infrared radiation, often appearing as "cold areas" in thermal images, requiring temperature correction using a reflection compensation model. Low-reflective materials (such as coal and wet clay) exhibit strong thermal radiation and are easily identifiable, but are susceptible to interference from dust and moisture, necessitating the introduction of anti-fouling algorithms to enhance stability. Temperature-sensitive materials (such as chemical raw materials and grains) are extremely sensitive to changes in ambient temperature, requiring a dynamic threshold adjustment mechanism to avoid misjudgments due to seasonal variations or changes in conveyor belt conditions.

[0018] To achieve precise adaptation to different materials, the detection system selects material type or learns infrared samples during the initialization phase, automatically calibrates feature extraction parameters and judgment thresholds, and establishes a material infrared characteristic database. The system can adaptively model the radiation, reflection and heat conduction behavior of specific transported materials, thereby achieving highly reliable detection with "one machine for multiple uses and one policy for one scenario" in multiple industries such as coal, metallurgy, and building materials.

[0019] Method B, Structural Reflection and Obstruction Handling: Specifically, the physical structure of the material inlet (such as metal liners, wear-resistant ceramic linings, inclined walls, etc.) can significantly interfere with infrared thermal imaging detection. These structural materials possess specific infrared reflection and radiation characteristics. For example, metal surfaces easily form high-reflection areas, while ceramic materials may produce non-uniform thermal images due to varying surface roughness. These fixed structures create continuous thermal patterns in the thermal image that are independent of the material's state, i.e., "structural thermal noise." To eliminate this interference, after initial installation, the detection system drives the conveyor belt equipment to run under no-load conditions, acquiring multiple frames of thermal image sequences under pure structural conditions. Statistical analysis is then used to generate a high-confidence background thermal image template. In actual material blockage monitoring, the detection system employs differential processing technology between real-time images and the background template to effectively remove the structural reflection components of the material inlet, thereby ensuring that subsequent feature extraction targets only the thermal behavior of the material itself.

[0020] For material discharge ports with complex structures, visual obstructions (such as internal support beams or inspection door frames), or large-angle inclinations, single-point infrared monitoring is prone to creating blind spots. To address this, the system supports a multi-sensor collaborative deployment scheme. By installing multiple infrared thermal imagers at different viewing angles (such as directly above, to the side, and in the opposite direction of flow), and performing temporal synchronization and spatial registration of images from each field of view, an image fusion algorithm is used to synthesize a complete panoramic temperature field of the material discharge port area. This scheme not only completely eliminates monitoring blind spots caused by structural obstructions but also cross-verifies the authenticity of suspected blockage areas through multi-angle data comparison, greatly improving the reliability and robustness of detection results in complex structural scenarios.

[0021] Method C, Algorithm Enhancement for Special Working Conditions: Specifically, to address the complex and variable environmental conditions in industrial settings, this system integrates multiple enhancement algorithms for special working conditions on top of the basic thermal imaging analysis framework to ensure detection reliability. In high-dust environments with high concentrations of dust, the detection system employs temporal filtering and dust penetration algorithms. By dynamically suppressing noise and correcting degradation models on continuous frames of thermal image data, it effectively improves the penetration of infrared radiation in dusty media and image stability. For scenarios involving the conveying of high-temperature materials (such as sintered ore and hot coke), due to the high base temperature of the materials themselves, the local temperature rise caused by material blockage is relatively insignificant. The detection system shifts its focus from absolute temperature thresholds to the gradient distribution analysis of the spatial temperature field, achieving early warning by identifying abnormal heat accumulation trends in localized areas within the material inlet.

[0022] Method B includes a method for modeling structural thermal noise, generating a background template, and removing the background. Specifically, it assumes that in the unloaded state of the material discharge port, only fixed structures such as metal lining plates, ceramic linings, and inclined walls exist. At this time, the infrared thermal image satisfies...

[0023] in It represents the temperature field dominated by the main structure of the discharge port, and does not include the thermal behavior of the material;

[0024] After initial installation or maintenance, the material blockage detection system performs no-load data acquisition to obtain continuous data. A frame of thermal image, expressed by a formula:

[0025]

[0026] Then construct the background mean template:

[0027]

[0028] Simultaneously calculate the background stability variance plot:

[0029]

[0030] Where B(x,y) is the structural background thermal image template, and V(x,y) is the structural stability confidence map; when V(x,y) is less than the threshold At that time, the pixel is considered a high-confidence structural region;

[0031] Assuming that the real-time thermal images collected during actual operation of the material blockage detection system are as follows: The process of constructing a differential image to remove the reflective components of the material outlet structure can be expressed by the following formula:

[0032] .

[0033] Method C uses a filtering algorithm for high-dust environments. Specifically, it assumes that the infrared characteristic energy of the dust medium in a high-dust environment can be approximated as an infrared radiation attenuation model, and employs a time-series degradation compensation filter, expressed by the following formula:

[0034]

[0035] in This is the filtered thermal image. Mixing coefficient The determination is based on the actual on-site conditions of the high-dust environment.

[0036] Method C also uses a gradient-dominated model for high-temperature material scenarios. Specifically, it assumes that when the material's base temperature is high, the overall material temperature has exceeded the blockage threshold, and therefore the absolute temperature rise characteristic is not significant. The blockage detection system defines the local temperature deviation by detecting the local heat accumulation trend, expressed by the formula:

[0037]

[0038] Local temperature L represents the neighborhood;

[0039] If consecutively satisfy If so, it is determined that there is an abnormal heat accumulation trend, among which This is the threshold for abnormal heat accumulation.

[0040] The early warning execution unit triggers the audible and visual alarm based on the analysis results of the intelligent computing unit, and outputs a linkage control signal to the central control system, which then links the external DCS system to perform actions such as stopping the conveyor or bypassing materials.

[0041] The intelligent computing unit includes a signal processing unit and an intelligent analysis unit, and the detection system also includes an early warning execution unit;

[0042] The signal processing unit and intelligent analysis unit are used to preprocess thermal imaging data, divide regions of interest (ROI), extract spatiotemporal thermal image features, and perform multi-feature fusion judgment. Based on the multi-feature fusion diagnostic model, the blockage status is determined to form a blockage identification and analysis result.

[0043] The multi-feature fusion diagnostic model is obtained by jointly modeling the spatial distribution features and temporal evolution features of infrared thermal image sequences, replacing the traditional single-point threshold judgment, and realizing intelligent identification of the material blockage state at the material discharge port.

[0044] The multi-feature fusion diagnostic model is obtained through the following steps:

[0045] Step 1: Image preprocessing and ROI segmentation; specifically:

[0046] Since the original infrared image may be affected by environmental noise, dust, etc., the signal processing unit first performs preprocessing on the image, including noise filtering, temperature calibration and other operations, to convert the original infrared signal into temperature field data.

[0047] Let the temperature field after pretreatment be expressed as Based on this, according to the fixed position of the material discharge port in the image or through initial calibration, the Region of Interest (ROI) is divided and denoted as the region set. ,like This indicates that the pixel is located within the material discharge port monitoring area;

[0048] Step 2: Multi-dimensional spatiotemporal feature extraction; specifically:

[0049] (1) Extracting spatial distribution features: First, extract features that reflect the spatial distribution of temperature;

[0050] Let the average temperature characteristic be expressed by the formula:

[0051] in: : Number of pixels within the ROI; :time average temperature

[0052] The area characteristics of the high-temperature region can be expressed by the formula:

[0053] in: High temperature threshold; : Indicator function, returns 1 if the condition is true, otherwise returns 0;

[0054] This feature is used to characterize the spatial extent of abnormally high temperature regions;

[0055] (2) Extracting temporal variation features: In order to describe the variation law of temperature field over time, the following features are introduced:

[0056] Temperature change rate characteristics:

[0057]

[0058] Temperature stability characteristics:

[0059] in: : Length of time window; Average temperature within the window;

[0060] The temporal variation characteristics are used to characterize the slowing down or even stopping of material flow when blockage occurs, and the temperature field often exhibits the characteristics of "local high temperature + time stability".

[0061] Step 3: Establish an intelligent diagnostic model based on multi-feature fusion, specifically as follows:

[0062] After extracting the spatial distribution features and temporal variation features mentioned above, each feature is normalized, and a comprehensive judgment function is constructed, expressed by the formula:

[0063] in: Features after normalization; Feature weights;

[0064] when When this happens, the system determines the current state to be either a blockage or a risk of blockage. The blockage threshold is set based on experience.

[0065] In step three, the definition and acquisition methods for normalized features and weight parameters are as follows: To eliminate differences in the dimensions, value ranges, and orders of magnitude of different features, and to ensure the stability and comparability of the multi-feature fusion judgment model, each feature is first normalized before fusing spatial distribution features and temporal variation features. Specific methods include:

[0066] Method 1. Definition and acquisition of normalized features: Assume that the normalized features include the normalized features of the high-temperature region area. Average temperature normalization characteristics and temperature stability normalization characteristics Its specific definition and calculation method are as follows:

[0067] Calculate the area normalization characteristics of high-temperature regions Specifically, let's assume at time... At that time, the area characteristics of the high-temperature region within the region of interest are as follows: The calculation method adopts the high-temperature region area feature calculation formula in step two. To facilitate the integration of the high-temperature region area normalization feature with other features, it is normalized to a dimensionless feature, defined as:

[0068]

[0069] in:

[0070] The minimum area of ​​the high-temperature zone obtained during system operation under normal material feeding conditions or without material blockage.

[0071] The maximum area of ​​the high-temperature zone obtained from historical data or manually calibrated blockage conditions;

[0072] and The data is obtained through historical data statistics during the initial operation phase of the blockage detection system, or determined by manual calibration during the on-site commissioning phase.

[0073] Calculate the normalized characteristics of the average temperature Specifically:

[0074] Set at time At that time, the average temperature within the ROI region was Its normalized expression is defined as:

[0075] in:

[0076] The lower limit of the average temperature collected by the system under normal material feeding and no material blockage conditions;

[0077] The upper limit of the average temperature collected under conditions of material blockage or near-blockage.

[0078] By using the above methods, the direct influence of changes in ambient temperature and differences in the initial temperature of materials on the judgment results can be eliminated, so that the temperature characteristics reflect the abnormal trend of the thermal field more than the absolute value.

[0079] Calculate the normalized characteristics of temperature stability Specifically: Let the length be... The temperature stability index calculated within the time window is To characterize the stability of temperature changes over time, its normalized expression is defined by the formula:

[0080] in:

[0081] The minimum temperature fluctuation obtained statistically under conditions of continuous material flow and no material blockage;

[0082] The maximum temperature fluctuation obtained statistically under conditions of material blockage or near-static accumulation;

[0083] In practical applications, the blockage state usually corresponds to a small temperature fluctuation. Therefore, this feature, when used in combination with other features, helps to improve the reliability of the judgment.

[0084] Method 2, the definition and acquisition method of feature weight parameters, specifically: In the multi-feature fusion judgment model, weight parameters are introduced. Used to describe the relative importance of different features in the determination of material blockage;

[0085] The meaning of the weight parameters is as follows:

[0086] Normalized characteristics of high-temperature region area The weighting coefficient is used to reflect the influence of the degree of expansion of the abnormal high temperature zone on the determination of blockage.

[0087] Mean temperature normalization characteristics The weighting coefficient is used to reflect the influence of the overall heat accumulation level on the determination of blockage.

[0088] Temperature stability normalization characteristics The weighting coefficients are used to reflect the influence of the thermal field time evolution characteristics on the blockage determination.

[0089] All introduced weight coefficients are non-negative real numbers and satisfy the following:

[0090] .

[0091] This invention comprehensively utilizes infrared thermal imaging sensing technology, image processing technology, and intelligent analysis algorithms to perform non-contact infrared light monitoring of the material operation status in belt conveyors, chutes, or material discharge areas. It is suitable for detecting and controlling material blockages in bulk materials such as coal, ore, sand, gravel, and metallurgical raw materials in industrial environments with high dust, high noise, and high vibration.

[0092] The detection system of this invention, based on the extraction and fusion of spatiotemporal thermal imaging features, further combines the infrared characteristics of different materials with the reflection features of the material discharge port structure to achieve precise adaptation and optimization for various industrial scenarios, thereby improving the practicality and robustness of the system.

[0093] To achieve precise adaptation to the infrared light characteristics of different materials, this invention supports material type selection or infrared sample learning during the initialization phase, and automatically calibrates feature extraction parameters and judgment thresholds. By establishing a material infrared characteristic database, the system can adaptively model the radiation, reflection, and heat conduction behavior of specific transported materials, thereby achieving highly reliable detection with "one machine for multiple uses and one strategy for one scenario" in multiple industries such as coal, metallurgy, and building materials.

[0094] To avoid interference from the physical structure of the material inlet (such as metal linings, wear-resistant ceramic liners, inclined walls, etc.) with infrared thermal imaging detection, this invention's system, after initial installation, drives the equipment to run in an unloaded state, acquiring multiple frames of thermal image sequences under pure structural conditions. High-confidence background thermal image templates are then generated through statistical analysis. In actual monitoring, the system employs differential processing technology between real-time images and the background template to effectively remove structural reflective components, thereby ensuring that subsequent feature extraction targets only the thermal behavior of the material itself.

[0095] This invention addresses the problem of blind spots in single-point infrared monitoring caused by complex structures, visual obstructions (such as internal support beams or inspection door frames), or steeply tilted material discharge ports. The system supports a multi-sensor collaborative deployment scheme. By installing multiple infrared thermal imagers at different viewing angles (e.g., directly above, to the side, and in the opposite direction of flow), and performing temporal synchronization and spatial registration of images from each field of view, an image fusion algorithm is used to synthesize a complete panoramic temperature field of the material discharge port area. This scheme not only completely eliminates monitoring blind spots caused by structural obstructions but also cross-verifies the authenticity of suspected blockage areas through multi-angle data comparison, greatly improving the reliability and robustness of detection results in complex structural scenarios.

[0096] This invention addresses the complex and ever-changing environmental conditions of industrial sites. The detection system, built upon a basic thermal imaging analysis framework, integrates multiple enhancement algorithms tailored to specific working conditions to ensure detection reliability. In high-dust environments, the system employs temporal filtering and dust penetration algorithms. By dynamically suppressing noise and correcting degradation models on consecutive frames of thermal image data, it effectively improves the penetration of infrared radiation in dusty media and enhances image stability.

[0097] For scenarios involving the conveying of high-temperature materials (such as sintered ore and hot coke), this invention addresses the characteristics of such materials having high base temperatures and relatively insignificant local temperature rises caused by material blockage. The detection system shifts its focus from absolute temperature thresholds to the gradient distribution analysis of the spatial temperature field, achieving early warning by identifying abnormal heat accumulation trends in local areas within the material inlet.

[0098] The present invention provides a material blockage detection system for the material discharge port based on infrared thermal imaging. It is a typical hardware and software collaborative system. It acquires the two-dimensional temperature field distribution information of the material discharge port area in real time through the infrared image acquisition unit. The signal processing and intelligent analysis unit further performs joint analysis on the spatial distribution characteristics and temporal evolution characteristics of the temperature field. Finally, the early warning execution unit outputs an alarm signal or linkage control command.

[0099] Unlike traditional blockage detection methods that rely solely on single-point temperature thresholds or mechanical contact sensors, this invention does not depend on whether a single temperature measurement point exceeds a set threshold. Instead, it comprehensively analyzes the infrared thermal image sequence to make a judgment on the blockage status from multiple dimensions, such as "how the temperature is distributed," "whether heat continues to accumulate in local areas," and "whether the temperature field changes significantly over time." This significantly improves the reliability and anti-interference capability of the detection. Attached Figure Description

[0100] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0101] Appendix Figure 1 This is a schematic diagram of the overall system principle structure in an embodiment of the present invention;

[0102] Appendix Figure 2 This is a schematic diagram of the installation of the thermal imaging device at the material discharge port in an embodiment of the present invention;

[0103] Appendix Figure 3 This is a schematic diagram of the internal structure of the thermal imaging device in an embodiment of the present invention;

[0104] Appendix Figure 4 This is a schematic diagram of a thermal image acquired by a long-wave infrared detector in an embodiment of the present invention;

[0105] Appendix Figure 5This is a flowchart illustrating the working process of the present invention; Detailed Implementation

[0106] Example:

[0107] As shown in the figure, a material blockage detection system based on infrared thermal imaging can perform material blockage detection by measuring the movement state of the bulk material at the material inlet with infrared light in a high-concentration dust environment. The detection system includes an infrared image acquisition unit and an intelligent computing unit.

[0108] The material discharge port is adjacent to the end of the material conveyor belt. The thermal imaging device of the infrared image acquisition unit includes a long-wave infrared detector installed at the material discharge port via a mounting bracket. The working wavelength range of the long-wave infrared detector is set to 8-14μm to acquire continuous infrared image sequences in high-concentration dust environments. This is to adapt to the environment of large dust concentrations and difficult visible light imaging in industrial sites where bulk material is transported. The field of view of the long-wave infrared detector is adjusted to cover the entire material discharge port area and its nearby key parts, so that the abnormal heat zone formed when the material discharge port is blocked can be fully included in the imaging range.

[0109] When the detection system is working, the long-wave infrared detector continuously outputs an infrared image sequence corresponding to the two-dimensional temperature field distribution at the material discharge port to the intelligent computing unit, denoted as I(x,y,t), where x and y represent the pixel coordinates in the image and t represents the time series index. First, the two-dimensional temperature field distribution information at the material discharge port is measured by the long-wave infrared detector to generate an infrared thermal image sequence. The intelligent computing unit analyzes the temperature gradient, hot zone area, temperature mean or fluctuation of the infrared thermal image sequence to obtain the temperature distribution at the material discharge port, the trend of continuous heat accumulation in local areas, and the trend of temperature field changes over time. It also performs joint analysis on the spatial distribution characteristics and temporal evolution characteristics of the temperature field, and adopts a multi-feature fusion diagnostic model that combines rule judgment and intelligent algorithms to accurately identify the blockage state at the material discharge port. Finally, the early warning execution unit outputs an alarm signal or linkage control command.

[0110] The long-wave infrared detector is installed on one side of the outer casing of the thermal imaging equipment, and a dust cover is provided at the detection end of the long-wave infrared detector.

[0111] The intelligent computing unit includes an embedded platform with AI computing capabilities, used to process the calculation process of the multi-feature fusion diagnostic model. The intelligent computing unit is installed on the other side of the device casing. The external interfaces of the intelligent computing unit include a power input and an RJ45 interface.

[0112] The thermal imaging equipment is mechanically fixed above or to the side of the material inlet with a metal bracket. The outer shell of the thermal imaging equipment is a dustproof, shockproof, and high-temperature resistant industrial protective shell.

[0113] If the infrared radiation and reflection characteristics of different materials or discharge port structures vary significantly when the material conveyor belt is conveying different bulk materials, and this directly and seriously affects the accuracy and reliability of the infrared light intensity detection of the discharge port blockage detection system, then the multi-feature fusion diagnostic model of the detection system, based on the extraction and fusion of spatiotemporal thermal imaging features, further combines the infrared characteristics of different bulk materials with the reflection characteristics of the discharge port structure to achieve precise adaptation and optimization for various industrial scenarios, thereby improving the system's practicality and robustness; the measurement, calibration, and sampling data calculation of the infrared light intensity distribution and changes at the discharge port are performed, specifically including:

[0114] Method A, Material Infrared Characteristic Classification and Feature Calibration, specifically addresses the significant differences in infrared radiation and reflection characteristics among different materials during bulk material transport, directly impacting the accuracy and reliability of thermal imaging detection. For instance, highly reflective materials (such as metal ores and dry sand) readily reflect ambient infrared radiation, often appearing as "cold areas" in thermal images, requiring temperature correction using a reflection compensation model. Low-reflective materials (such as coal and wet clay) exhibit strong thermal radiation and are easily identifiable, but are susceptible to interference from dust and moisture, necessitating the introduction of anti-fouling algorithms to enhance stability. Temperature-sensitive materials (such as chemical raw materials and grains) are extremely sensitive to changes in ambient temperature, requiring a dynamic threshold adjustment mechanism to avoid misjudgments due to seasonal variations or changes in conveyor belt conditions.

[0115] To achieve precise adaptation to different materials, the detection system selects material type or learns infrared samples during the initialization phase, automatically calibrates feature extraction parameters and judgment thresholds, and establishes a material infrared characteristic database. The system can adaptively model the radiation, reflection and heat conduction behavior of specific transported materials, thereby achieving highly reliable detection with "one machine for multiple uses and one policy for one scenario" in multiple industries such as coal, metallurgy, and building materials.

[0116] To facilitate on-site debugging and adaptation, the system provides the following configurable parameter modules:

[0117] Table 1 shows the system's configurable parameters and their adaptation to different scenarios.

[0118] Table 1: System Configurable Parameters and Scenario Adaptation Table

[0119]

[0120] Method B, Structural Reflection and Obstruction Handling: Specifically, the physical structure of the material inlet (such as metal liners, wear-resistant ceramic linings, inclined walls, etc.) can significantly interfere with infrared thermal imaging detection. These structural materials possess specific infrared reflection and radiation characteristics. For example, metal surfaces easily form high-reflection areas, while ceramic materials may produce non-uniform thermal images due to varying surface roughness. These fixed structures create continuous thermal patterns in the thermal image that are independent of the material's state, i.e., "structural thermal noise." To eliminate this interference, after initial installation, the detection system drives the conveyor belt equipment to run under no-load conditions, acquiring multiple frames of thermal image sequences under pure structural conditions. Statistical analysis is then used to generate a high-confidence background thermal image template. In actual material blockage monitoring, the detection system employs differential processing technology between real-time images and the background template to effectively remove the structural reflection components of the material inlet, thereby ensuring that subsequent feature extraction targets only the thermal behavior of the material itself.

[0121] For material discharge ports with complex structures, visual obstructions (such as internal support beams or inspection door frames), or large-angle inclinations, single-point infrared monitoring is prone to creating blind spots. To address this, the system supports a multi-sensor collaborative deployment scheme. By installing multiple infrared thermal imagers at different viewing angles (such as directly above, to the side, and in the opposite direction of flow), and performing temporal synchronization and spatial registration of images from each field of view, an image fusion algorithm is used to synthesize a complete panoramic temperature field of the material discharge port area. This scheme not only completely eliminates monitoring blind spots caused by structural obstructions but also cross-verifies the authenticity of suspected blockage areas through multi-angle data comparison, greatly improving the reliability and robustness of detection results in complex structural scenarios.

[0122] Method C, Algorithm Enhancement for Special Working Conditions: Specifically, to address the complex and variable environmental conditions in industrial settings, this system integrates multiple enhancement algorithms for special working conditions on top of the basic thermal imaging analysis framework to ensure detection reliability. In high-dust environments with high concentrations of dust, the detection system employs temporal filtering and dust penetration algorithms. By dynamically suppressing noise and correcting degradation models on continuous frames of thermal image data, it effectively improves the penetration of infrared radiation in dusty media and image stability. For scenarios involving the conveying of high-temperature materials (such as sintered ore and hot coke), due to the high base temperature of the materials themselves, the local temperature rise caused by material blockage is relatively insignificant. The detection system shifts its focus from absolute temperature thresholds to the gradient distribution analysis of the spatial temperature field, achieving early warning by identifying abnormal heat accumulation trends in localized areas within the material inlet.

[0123] Method B includes a method for modeling structural thermal noise, generating a background template, and removing the background. Specifically, it assumes that in the unloaded state of the material discharge port, only fixed structures such as metal lining plates, ceramic linings, and inclined walls exist. At this time, the infrared thermal image satisfies...

[0124] in It represents the temperature field dominated by the main structure of the discharge port, and does not include the thermal behavior of the material;

[0125] After initial installation or maintenance, the material blockage detection system performs no-load data acquisition to obtain continuous data. A frame of thermal image, expressed by a formula:

[0126]

[0127] Then construct the background mean template:

[0128]

[0129] Simultaneously calculate the background stability variance plot:

[0130]

[0131] Where B(x,y) is the structural background thermal image template, and V(x,y) is the structural stability confidence map; when V(x,y) is less than the threshold At that time, the pixel is considered a high-confidence structural region;

[0132] Assuming that the real-time thermal images collected during actual operation of the material blockage detection system are as follows: The process of constructing a differential image to remove the reflective components of the material outlet structure can be expressed by the following formula:

[0133] .

[0134] Method C uses a filtering algorithm for high-dust environments. Specifically, it assumes that the infrared characteristic energy of the dust medium in a high-dust environment can be approximated as an infrared radiation attenuation model, and employs a time-series degradation compensation filter, expressed by the following formula:

[0135]

[0136] in This is the filtered thermal image. Mixing coefficient The determination is based on the actual on-site conditions of the high-dust environment.

[0137] Method C also uses a gradient-dominated model for high-temperature material scenarios. Specifically, it assumes that when the material's base temperature is high, the overall material temperature has exceeded the blockage threshold, and therefore the absolute temperature rise characteristic is not significant. The blockage detection system defines the local temperature deviation by detecting the local heat accumulation trend, expressed by the formula:

[0138]

[0139] Local temperature L represents the neighborhood;

[0140] If consecutively satisfy If so, it is determined that there is an abnormal heat accumulation trend, among which This is the threshold for abnormal heat accumulation.

[0141] The early warning execution unit triggers the audible and visual alarm based on the analysis results of the intelligent computing unit, and outputs a linkage control signal to the central control system, which then links the external DCS system to perform actions such as stopping the conveyor or bypassing materials.

[0142] The intelligent computing unit includes a signal processing unit and an intelligent analysis unit, and the detection system also includes an early warning execution unit;

[0143] The signal processing unit and intelligent analysis unit are used to preprocess thermal imaging data, divide regions of interest (ROI), extract spatiotemporal thermal image features, and perform multi-feature fusion judgment. Based on the multi-feature fusion diagnostic model, the blockage status is determined to form a blockage identification and analysis result.

[0144] The multi-feature fusion diagnostic model is obtained by jointly modeling the spatial distribution features and temporal evolution features of infrared thermal image sequences, replacing the traditional single-point threshold judgment, and realizing intelligent identification of the material blockage state at the material discharge port.

[0145] The multi-feature fusion diagnostic model is obtained through the following steps:

[0146] Step 1: Image preprocessing and ROI segmentation; specifically:

[0147] Since the original infrared image may be affected by environmental noise, dust, etc., the signal processing unit first performs preprocessing on the image, including noise filtering, temperature calibration and other operations, to convert the original infrared signal into temperature field data.

[0148] Let the temperature field after pretreatment be expressed as Based on this, according to the fixed position of the material discharge port in the image or through initial calibration, the Region of Interest (ROI) is divided and denoted as the region set. ,like This indicates that the pixel is located within the material discharge port monitoring area;

[0149] Step 2: Multi-dimensional spatiotemporal feature extraction; specifically:

[0150] (1) Extracting spatial distribution features: First, extract features that reflect the spatial distribution of temperature;

[0151] Let the average temperature characteristic be expressed by the formula:

[0152] in: : Number of pixels within the ROI; :time average temperature

[0153] The area characteristics of the high-temperature region can be expressed by the formula:

[0154] in: High temperature threshold; : Indicator function, returns 1 if the condition is true, otherwise returns 0;

[0155] This feature is used to characterize the spatial extent of abnormally high temperature regions;

[0156] (2) Extracting temporal variation features: In order to describe the variation law of temperature field over time, the following features are introduced:

[0157] Temperature change rate characteristics:

[0158]

[0159] Temperature stability characteristics:

[0160] in: : Length of time window; Average temperature within the window;

[0161] The temporal variation characteristics are used to characterize the slowing down or even stopping of material flow when blockage occurs, and the temperature field often exhibits the characteristics of "local high temperature + time stability".

[0162] Step 3: Establish an intelligent diagnostic model based on multi-feature fusion, specifically as follows:

[0163] After extracting the spatial distribution features and temporal variation features mentioned above, each feature is normalized, and a comprehensive judgment function is constructed, expressed by the formula:

[0164] in: Features after normalization; Feature weights;

[0165] when When this happens, the system determines the current state to be either a blockage or a risk of blockage. The blockage threshold is set based on experience.

[0166] In step three, the definition and acquisition methods for normalized features and weight parameters are as follows: To eliminate differences in the dimensions, value ranges, and orders of magnitude of different features, and to ensure the stability and comparability of the multi-feature fusion judgment model, each feature is first normalized before fusing spatial distribution features and temporal variation features. Specific methods include:

[0167] Method 1. Definition and acquisition of normalized features: Assume that the normalized features include the normalized features of the high-temperature region area. Average temperature normalization characteristics and temperature stability normalization characteristics Its specific definition and calculation method are as follows:

[0168] Calculate the area normalization characteristics of high-temperature regions Specifically, let's assume at time... At that time, the area characteristics of the high-temperature region within the region of interest are as follows: The calculation method adopts the high-temperature region area feature calculation formula in step two. To facilitate the integration of the high-temperature region area normalization feature with other features, it is normalized to a dimensionless feature, defined as:

[0169]

[0170] in:

[0171] The minimum area of ​​the high-temperature zone obtained during system operation under normal material feeding conditions or without material blockage.

[0172] The maximum area of ​​the high-temperature zone obtained from historical data or manually calibrated blockage conditions;

[0173] and The data is obtained through historical data statistics during the initial operation phase of the blockage detection system, or determined by manual calibration during the on-site commissioning phase.

[0174] Calculate the normalized characteristics of the average temperature Specifically:

[0175] Set at time At that time, the average temperature within the ROI region was Its normalized expression is defined as:

[0176] in:

[0177] The lower limit of the average temperature collected by the system under normal material feeding and no material blockage conditions;

[0178] The upper limit of the average temperature collected under conditions of material blockage or near-blockage.

[0179] By using the above methods, the direct influence of changes in ambient temperature and differences in the initial temperature of materials on the judgment results can be eliminated, so that the temperature characteristics reflect the abnormal trend of the thermal field more than the absolute value.

[0180] Calculate the normalized characteristics of temperature stability Specifically: Let the length be... The temperature stability index calculated within the time window is To characterize the stability of temperature changes over time, its normalized expression is defined by the formula:

[0181] in:

[0182] The minimum temperature fluctuation obtained statistically under conditions of continuous material flow and no material blockage;

[0183] The maximum temperature fluctuation obtained statistically under conditions of material blockage or near-static accumulation;

[0184] In practical applications, the blockage state usually corresponds to a small temperature fluctuation. Therefore, this feature, when used in combination with other features, helps to improve the reliability of the judgment.

[0185] Method 2, the definition and acquisition method of feature weight parameters, specifically: In the multi-feature fusion judgment model, weight parameters are introduced. Used to describe the relative importance of different features in the determination of material blockage;

[0186] The meaning of the weight parameters is as follows:

[0187] Normalized characteristics of high-temperature region area The weighting coefficient is used to reflect the influence of the degree of expansion of the abnormal high temperature zone on the determination of blockage.

[0188] Mean temperature normalization characteristics The weighting coefficient is used to reflect the influence of the overall heat accumulation level on the determination of blockage.

[0189] Temperature stability normalization characteristics The weighting coefficients are used to reflect the influence of the thermal field time evolution characteristics on the blockage determination.

[0190] All introduced weight coefficients are non-negative real numbers and satisfy the following:

[0191] .

[0192] This example presents a material blockage detection system for the material discharge port based on infrared thermal imaging. For example... Figure 1As shown, the system mainly includes an infrared image acquisition unit, a signal processing and intelligent analysis unit, and an early warning execution unit. The infrared image acquisition unit uses an infrared thermal imager installed at the material inlet to acquire real-time temperature distribution images of the material inlet and surrounding area. The signal processing and intelligent analysis unit preprocesses the acquired thermal image data, performs ROI segmentation, extracts spatiotemporal thermal image features, and performs multi-feature fusion judgment to form a blockage identification result. The early warning execution unit triggers an audible and visual alarm based on the analysis result and links the DCS system to execute actions such as conveyor shutdown or material bypass. The core algorithm of this invention extracts the spatiotemporal features of the thermal image (such as temperature gradient, hot zone area, average temperature, or temperature fluctuation), and adopts a multi-feature fusion diagnostic model combining rule judgment and intelligent algorithms to accurately identify the blockage status of the material inlet, significantly reducing the false alarm rate and providing early warning.

Claims

1. A material blockage detection system for a conveyor belt discharge port based on infrared thermal imaging, capable of accurately measuring and acquiring infrared light intensity data of the movement of bulk materials at the discharge port in a high-concentration dust environment, characterized in that: The detection system includes an infrared image acquisition unit and an intelligent computing unit; The material discharge port is adjacent to the end of the material conveyor belt. The thermal imaging device of the infrared image acquisition unit includes a long-wave infrared detector installed at the material discharge port via a mounting bracket. The working band range of the long-wave infrared detector is set to 8-14μm to acquire a continuous infrared image sequence in a high-concentration dust environment. Its field of view can allow the abnormal hot zone formed when the material discharge port is blocked to be completely included in the imaging range of the thermal imaging device. When the detection system is working, the long-wave infrared detector continuously outputs an infrared image sequence corresponding to the temperature field distribution at the material discharge port to the intelligent computing unit, denoted as I(x,y,t), where x and y represent the pixel coordinates in the image and t represents the time series index. The intelligent computing unit analyzes the temperature gradient, hot zone area, and average or fluctuating temperature of the infrared thermal image sequence to obtain the temperature distribution at the material discharge port, the trend of continuous heat accumulation in local areas, and the trend of temperature field changes over time. It also analyzes the spatial distribution characteristics and temporal evolution characteristics of the temperature field and uses a multi-feature fusion diagnostic model to determine whether the material discharge port is blocked.

2. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 1, characterized in that: The long-wave infrared detector is installed on one side of the outer casing of the thermal imaging equipment, and a dust cover is provided at the detection end of the long-wave infrared detector. The intelligent computing unit includes an embedded platform for processing the calculation process of the multi-feature fusion diagnostic model. The intelligent computing unit is installed on the other side of the device housing. The external interfaces of the intelligent computing unit include a power input and an RJ45 interface. The thermal imaging equipment is mechanically fixed above or to the side of the material inlet with a metal bracket. The outer shell of the thermal imaging equipment is a dustproof, shockproof, and high-temperature resistant industrial protective shell.

3. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 1, characterized in that: If the infrared radiation and reflection characteristics of different materials or discharge port structures vary greatly when the material conveyor belt is conveying different bulk materials, then the multi-feature fusion diagnostic model of the detection system will further combine the infrared characteristics of different bulk materials and the reflection characteristics of the discharge port structure on the basis of extracting and fusing spatiotemporal thermal image features. The measurement, calibration, and calculation of infrared light intensity distribution and changes at the material discharge port are performed. Specific methods include: Method A, material infrared characteristic classification and feature calibration, specifically: for high reflectivity materials, temperature correction is performed by combining a reflection compensation model; for low reflectivity materials, an anti-fouling algorithm is introduced to enhance stability; and for temperature-sensitive materials, a dynamic threshold adjustment mechanism is adopted to avoid misjudgment due to seasonal changes or changes in the working conditions of the material conveyor belt. During the initialization phase, the material blockage detection system selects the material type or learns infrared samples, automatically calibrates the feature extraction parameters and judgment thresholds, establishes a material infrared characteristic database, adaptively models the radiation, reflection and heat conduction behavior of specific conveyed materials, establishes a configuration parameter and scenario adaptation association data table, and performs "one machine for multiple uses, one policy for one scenario" material blockage detection at the material drop-out scene of conveyor belt equipment in the coal, metallurgy and building materials industries. Method B, structural reflection and occlusion processing, specifically involves: after initial installation, the detection system drives the conveyor belt equipment to run under no-load conditions, collecting multiple frames of thermal image sequences under pure structural conditions, and generating a high-confidence background thermal image template through statistical analysis; in actual material blockage monitoring, the detection system uses differential processing technology between real-time images and background templates to remove structural reflection components at the material discharge port, ensuring that subsequent feature extraction is only for the thermal behavior of the material itself; For material discharge port structures that are complex, have visual obstructions, or are tilted at large angles, multiple infrared thermal imagers are installed at different viewing angles, and the images of each field of view are synchronized in time and registered in space. Then, an image fusion algorithm is used to synthesize a complete panoramic temperature field of the material discharge port area. Method C, algorithm enhancement under special working conditions, specifically: In high-dust environments with high concentrations of dust, the detection system adopts time-series filtering and dust penetration algorithms. By performing dynamic noise suppression and degradation model correction on continuous frame thermal image data, the penetration of infrared radiation in dusty media and image stability are effectively improved. For scenarios involving the conveying of high-temperature materials, the detection system shifts its focus from absolute temperature thresholds to gradient distribution analysis of the spatial temperature field, and provides early warning by identifying abnormal heat accumulation trends in local areas within the material inlet.

4. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 3, characterized in that: Method B includes a method for modeling structural thermal noise, generating a background template, and removing the background. Specifically, it assumes that in the unloaded state of the material discharge port, only fixed structures such as metal lining plates, ceramic linings, and inclined walls exist. At this time, the infrared thermal image satisfies... in It represents the temperature field dominated by the main structure of the discharge port, and does not include the thermal behavior of the material; After initial installation or maintenance, the material blockage detection system performs no-load data acquisition to obtain continuous data. A frame of thermal image, expressed by a formula: Then construct the background mean template: Simultaneously calculate the background stability variance plot: Where B(x,y) is the structural background thermal image template, and V(x,y) is the structural stability confidence map; when V(x,y) is less than the threshold At that time, the pixel is considered a high-confidence structural region; Assuming that the real-time thermal images collected during actual operation of the material blockage detection system are as follows: The process of constructing a differential image to remove the reflective components of the material outlet structure can be expressed by the following formula: 。 5. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 3, characterized in that: Method C uses a filtering algorithm for high-dust environments. Specifically, it assumes that the infrared characteristic energy of the dust medium in a high-dust environment can be approximated as an infrared radiation attenuation model, and employs a time-series degradation compensation filter, expressed by the following formula: in This is the filtered thermal image. Mixing coefficient The determination is based on the actual on-site conditions of the high-dust environment.

6. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 3, characterized in that: Method C also uses a gradient-dominated model for high-temperature material scenarios. Specifically, it assumes that when the material's base temperature is high, the overall material temperature has exceeded the blockage threshold, and therefore the absolute temperature rise characteristic is not significant. The blockage detection system defines the local temperature deviation by detecting the local heat accumulation trend, expressed by the formula: Local temperature L represents the neighborhood; If consecutively satisfy If so, it is determined that there is an abnormal heat accumulation trend, among which This is the threshold for abnormal heat accumulation.

7. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 1, characterized in that: The early warning execution unit triggers the audible and visual alarm based on the analysis results of the intelligent computing unit, and outputs a linkage control signal to the central control system to link the external system to execute equipment actions such as stopping the conveyor or bypassing materials.

8. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 1, characterized in that: The intelligent computing unit includes a signal processing unit and an intelligent analysis unit, and the detection system also includes an early warning execution unit; The signal processing unit and intelligent analysis unit are used to preprocess thermal imaging data, divide regions of interest (ROI), extract spatiotemporal thermal image features, and perform multi-feature fusion judgment. Based on the multi-feature fusion diagnostic model, the blockage status is determined to form a blockage identification and analysis result. The multi-feature fusion diagnostic model is obtained by jointly modeling the spatial distribution features and temporal evolution features of infrared thermographic sequences.

9. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 4, characterized in that: The multi-feature fusion diagnostic model is obtained through the following steps: Step 1: Image preprocessing and ROI segmentation; specifically: The signal processing unit first preprocesses the image, including noise filtering and temperature calibration, converting the raw infrared signal into temperature field data; Let the temperature field after pretreatment be expressed as Based on this, according to the fixed position of the discharge port in the image or through initial calibration, the Region of Interest (ROI) is divided and denoted as the region set. ,like This indicates that the pixel is located within the material discharge port monitoring area; Step 2: Multi-dimensional spatiotemporal feature extraction; specifically: (1) Extracting spatial distribution features: First, extract features that reflect the spatial distribution of temperature; Let the average temperature characteristic be expressed by the formula: in: : Number of pixels within the ROI; :time average temperature The area characteristics of the high-temperature region can be expressed by the formula: in: High temperature threshold; : Indicator function, returns 1 if the condition is true, otherwise returns 0; This feature is used to characterize the spatial extent of abnormally high temperature regions; (2) Extracting temporal variation features: In order to describe the variation law of temperature field over time, the following features are introduced: Temperature change rate characteristics: Temperature stability characteristics: in: : Length of time window; Average temperature within the window; The temporal variation characteristics are used to characterize the slowing down or even stopping of material flow when blockage occurs, and the temperature field exhibits the characteristics of "local high temperature + time stability"; Step 3: Establish an intelligent diagnostic model based on multi-feature fusion, specifically as follows: After extracting spatial distribution features and temporal variation features, normalization is performed, and a comprehensive judgment function is constructed, expressed by the formula: in: Features after normalization; Feature weights; when When this happens, the system determines the current state to be either a blockage or a risk of blockage. The blockage threshold is set based on experience.

10. The material blockage detection system for the discharge port based on infrared thermal imaging according to claim 9, characterized in that: In step three, the definition and acquisition methods for normalized features and weight parameters are as follows: To eliminate differences in the dimensions, value ranges, and orders of magnitude of different features, and to ensure the stability and comparability of the multi-feature fusion judgment model, each feature is first normalized before fusing spatial distribution features and temporal variation features. Specific methods include: Method 1. Definition and acquisition of normalized features: Assume that the normalized features include the normalized features of the high-temperature region area. Average temperature normalization characteristics and temperature stability normalization characteristics Its specific definition and calculation method are as follows: Calculate the area normalization characteristics of high-temperature regions Specifically, let's assume at time... At that time, the area characteristics of the high-temperature region within the region of interest are as follows: The calculation method adopts the high-temperature region area feature calculation formula in step two. To facilitate the integration of the high-temperature region area normalization feature with other features, it is normalized to a dimensionless feature, defined as: in: The minimum area of ​​the high-temperature zone obtained during system operation under normal material feeding conditions or without material blockage. The maximum area of ​​the high-temperature zone obtained from historical data or manually calibrated blockage conditions; and The data is obtained through historical data statistics during the initial operation phase of the blockage detection system, or determined by manual calibration during the on-site commissioning phase. Calculate the normalized characteristics of the average temperature Specifically: Set at time At that time, the average temperature within the ROI region was Its normalized expression is defined as: in: The lower limit of the average temperature collected by the system under normal material feeding and no material blockage conditions; The upper limit of the average temperature collected under conditions of material blockage or near-blockage. Calculate the normalized characteristics of temperature stability Specifically: Let the length be... The temperature stability index calculated within the time window is To characterize the stability of temperature changes over time, its normalized expression is defined by the formula: in: The minimum temperature fluctuation obtained statistically under conditions of continuous material flow and no material blockage; The maximum temperature fluctuation obtained statistically under conditions of material blockage or near-static accumulation; Method 2, the definition and acquisition method of feature weight parameters, specifically: In the multi-feature fusion judgment model, weight parameters are introduced. Used to describe the relative importance of different features in the determination of material blockage; The meaning of the weight parameters is as follows: Normalized characteristics of high-temperature region area The weighting coefficient is used to reflect the influence of the degree of expansion of the abnormal high temperature zone on the determination of blockage. Mean temperature normalization characteristics The weighting coefficient is used to reflect the influence of the overall heat accumulation level on the determination of blockage. Temperature stability normalization characteristics The weighting coefficients are used to reflect the influence of the thermal field time evolution characteristics on the blockage determination. All introduced weight coefficients are non-negative real numbers and satisfy the following: 。

Images