Intelligent monitoring method and platform for coal surface moisture content based on infrared imaging

Abstract

The application discloses an intelligent coal surface moisture content monitoring method and platform based on infrared imaging, and relates to the technical field of moisture content monitoring.The method comprises the following steps: activating an infrared thermal imager to continuously monitor target coal under thermal excitation, and obtaining a target infrared image sequence; performing space-time feature analysis on the target infrared image sequence based on a predetermined thermal response feature, and obtaining a target thermal response feature parameter; and traversing and comparing the target thermal response feature parameter in a moisture content thermal response feature database, and obtaining a target moisture content.The technical problem that the existing coal surface moisture content detection relies on a contact type or manual sampling mode, resulting in low detection efficiency and difficulty in realizing online continuous monitoring, is solved, and the technical effect of improving the real-time performance, accuracy and anti-interference capability of coal surface moisture content monitoring by introducing a pulse type thermal excitation and combining space-time feature analysis of an infrared image sequence and database traversal comparison is achieved.

Description

Technical Field

[0001] This invention relates to the field of moisture content monitoring technology, specifically to an intelligent monitoring method and platform for coal surface moisture content based on infrared imaging. Background Technology

[0002] Coal, as a crucial energy source for industries such as power generation, metallurgy, chemicals, and building materials, has quality parameters that directly impact combustion efficiency, transportation safety, storage stability, and subsequent processing and utilization. Among these, coal surface moisture content is a key indicator of coal's condition. Excessive moisture content not only reduces the actual calorific value of coal and increases transportation and handling costs, but can also lead to problems such as blockage, adhesion, freezing, and clumping during storage, transportation, and use, thereby affecting the continuity and stability of the production system. Especially in application scenarios such as open-air stockpiles, coal washing workshops, coal conveying corridors, and port transshipment, the surface moisture content of coal exhibits strong dynamic changes due to factors such as rainfall, spraying, ambient humidity, and internal water accumulation. Therefore, rapid, accurate, and continuous monitoring of coal surface moisture content is of significant practical importance.

[0003] In existing technologies, methods for detecting the moisture content of coal mainly employ manual sampling, drying and weighing, capacitive testing, microwave testing, or near-infrared testing. While manual sampling combined with laboratory testing offers some accuracy, it typically suffers from long testing cycles, cumbersome operations, and the inability to provide real-time feedback on the field status, making it difficult to meet the needs of online monitoring in industrial settings. Although capacitive, microwave, or near-infrared testing methods improve efficiency to some extent, they are still susceptible to factors such as differences in coal type, particle size distribution, surface roughness, temperature fluctuations, dust interference, and the installation environment, resulting in insufficient stability of the test results. Summary of the Invention

[0004] This application provides an intelligent monitoring method and platform for coal surface moisture content based on infrared imaging. The key aspect lies in addressing the technical obstacle of accurately extracting stable moisture content characterization features and achieving high-precision online determination due to the dynamic changes in infrared image sequences in coal stockpiles and transportation scenarios, the significant spatiotemporal coupling characteristics of thermal responses, and the susceptibility to interference from environmental and coal type differences. This is achieved through spatiotemporal feature analysis of the multidimensional thermal response characteristics of coal under pulsed thermal excitation and optimization of a similarity matching algorithm based on a feature database. Combined with an integrated data processing flow encompassing infrared image sequence acquisition, feature extraction, database comparison, and result calibration, this method achieves non-contact, continuous, and intelligent monitoring of coal surface moisture content, improving detection accuracy and real-time performance.

[0005] The first aspect of this application provides a method for intelligent monitoring of coal surface moisture content based on infrared imaging, the method comprising: Step S100: Activate the infrared thermal imager to continuously monitor the target coal under thermal excitation and obtain a target infrared image sequence; Step S200: Perform spatiotemporal feature analysis on the target infrared image sequence based on predetermined thermal response characteristics to obtain target thermal response feature parameters; Step S300: Compare the target thermal response feature parameters in the moisture content thermal response feature database to obtain the target moisture content; wherein, the thermal excitation refers to pulsed thermal radiation with a predetermined pulse width and predetermined pulse intensity.

[0006] A second aspect of this application provides an intelligent monitoring platform for coal surface moisture content based on infrared imaging, the platform comprising: Continuous monitoring module: Activates the infrared thermal imager to continuously monitor the target coal under thermal excitation and obtains the target infrared image sequence; Feature analysis module: Performs spatiotemporal feature analysis on the target infrared image sequence based on predetermined thermal response features to obtain the target thermal response feature parameters; Feature comparison module: Traverses and compares the target thermal response feature parameters in the moisture content thermal response feature database to obtain the target moisture content.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, under pulsed thermal radiation with preset pulse width and intensity, controlled thermal excitation is applied to the target coal, and an infrared thermal imaging device is simultaneously activated to continuously acquire data on its surface temperature changes, thereby obtaining an infrared image sequence of the target coal under dynamic thermal action. Then, based on a pre-defined thermal response feature system, a comprehensive analysis of the temporal evolution and spatial distribution of the acquired infrared image sequence is performed to extract target thermal response feature parameters that characterize the temperature change pattern of the coal after heating. Finally, the target thermal response feature parameters are traversed, matched, and compared with standard features in a pre-constructed moisture content thermal response feature database to inversely obtain the corresponding moisture content of the target coal, achieving intelligent moisture content identification based on thermal response features. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic diagram of the process for intelligent monitoring of coal surface moisture content based on infrared imaging, provided in an embodiment of this application.

[0010] Figure 2A schematic diagram of the structure of an intelligent monitoring platform for coal surface moisture content based on infrared imaging, provided in an embodiment of this application.

[0011] Figure labeling: Continuous monitoring module 11, feature analysis module 12, feature comparison module 13. Detailed Implementation

[0012] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0013] Example 1, as Figure 1 As shown, this application provides an intelligent monitoring method for surface moisture content of coal based on infrared imaging. The method includes: Step S100: Activate the infrared thermal imager to continuously monitor the target coal under thermal excitation and obtain the target infrared image sequence; wherein, the thermal excitation refers to pulsed thermal radiation with a predetermined pulse width and predetermined pulse intensity.

[0014] In this embodiment, the target coal is first placed at a predetermined monitoring position, and the lens of the infrared thermal imager is directed towards the surface of the target coal to ensure that the entire target coal is within the effective field of view of the infrared thermal imager. Before monitoring, a thermal excitation device is activated to perform pulsed thermal radiation heating on the target coal. This pulsed thermal radiation refers to applying a short-duration thermal radiation energy input with a predetermined pulse width and predetermined pulse intensity to the surface of the target coal within a preset time window. The predetermined pulse width is set according to the coal particle size, accumulation thickness, initial surface temperature, and target detection accuracy. The predetermined pulse intensity is used to control the thermal radiation energy applied to the surface of the target coal per unit time, so that the surface of the target coal forms a recognizable transient temperature response after excitation, while avoiding surface ablation, thermal damage, or temperature saturation due to overheating. The thermal excitation device is a pulsed infrared radiator, a pulsed halogen lamp, a pulsed heating lamp array, or other devices capable of outputting controlled pulsed thermal radiation. Before the thermal excitation begins, the infrared thermal imager first acquires an initial background thermal image of the target coal as a benchmark for temperature rise analysis. During the thermal excitation process and within a predetermined cooling time after the thermal excitation ends, the infrared thermal imager continuously monitors the target coal according to a preset sampling frequency, acquiring multiple frames of infrared images corresponding to changes in the surface temperature field of the target coal in real time. This results in a target infrared image sequence that reflects the thermal response changes of the target coal during the heating and subsequent cooling stages. Each frame in this target infrared image sequence carries surface heat distribution information at the corresponding acquisition time and is stored and numbered in chronological order, providing a data basis for the subsequent extraction of thermal response characteristics related to coal moisture content.

[0015] Step S200: Perform spatiotemporal feature analysis on the target infrared image sequence based on predetermined thermal response characteristics to obtain target thermal response feature parameters.

[0016] In one embodiment, after receiving the target infrared image sequence acquired by the infrared thermal imager, each frame of the infrared image is first preprocessed, including non-uniformity correction and bad pixel replacement, to eliminate the interference of detector noise on temperature measurement accuracy. Non-uniformity correction can be performed using a two-point correction method, i.e., acquiring two reference images of a low-temperature blackbody and a high-temperature blackbody, calculating the gain coefficient and bias coefficient of each pixel, and thus linearly normalizing the original grayscale value. For detectors with strong nonlinear responses, a multi-point correction method can also be used for piecewise linear fitting. Bad pixel replacement can be performed using neighborhood interpolation or median filtering. Specifically, by searching a pre-calibrated bad pixel location table, for the coordinate position of the marked bad pixel, neighborhood interpolation is used to estimate the replacement value using the grayscale values ​​of surrounding effective pixels, or median filtering is directly used to take the median of the neighborhood for replacement, ensuring the continuity of the image sequence and the accuracy of spatiotemporal feature analysis. Subsequently, the temperature value of each pixel position in the target infrared image sequence is extracted over time to construct temperature-time response curves at various spatial locations on the target coal surface.

[0017] In the time dimension analysis, for the temperature-time response curve of each pixel, the turning point where the temperature changes from rising to falling after the pulsed thermal excitation is automatically identified by using the first-order difference method or a five-point cubic smoothing filter to find the maximum value. The time corresponding to this turning point is recorded as the peak time of the thermal response for that pixel, and the temperature value at that moment is recorded as the peak temperature for that pixel. In the cooling stage after the peak time, a data segment within the range where the temperature drops from the peak to 1 / e (approximately 36.8%) of the peak temperature or a predetermined proportion is selected. The logarithmic temperature-time relationship of this data segment is linearly fitted using the least squares method, and the absolute value of the slope of the fitted line is taken as the temperature field decay rate of that pixel. In the spatial dimension analysis, based on the peak temperature values ​​calculated for each pixel, a two-dimensional matrix of the peak temperature distribution on the target coal surface is generated, and the mean, standard deviation, and coefficient of variation of this matrix are calculated. The mean represents the overall thermal response level, and the ratio of the standard deviation to the mean is taken as the spatial uniformity of the peak temperature distribution, which characterizes the degree of difference in heat diffusion caused by uneven moisture content distribution. To further reflect the anisotropic characteristics of thermal diffusion, a two-dimensional Fourier transform or gray-level co-occurrence matrix texture analysis is performed on the peak temperature distribution matrix to extract texture features such as energy, entropy, or contrast as auxiliary spatial feature parameters. Finally, the peak thermal response time, temperature field decay rate, peak temperature spatial distribution uniformity, and their statistical characteristic parameters are integrated to construct a multidimensional target thermal response feature parameter for characterizing the current moisture content of the target coal. This facilitates subsequent similarity comparison and moisture content determination in the moisture content thermal response feature database. Through this method, the thermal change information evolving over time in the target infrared image sequence can be jointly characterized with the surface spatial temperature distribution information, thereby improving the accuracy and stability of identifying the surface moisture content of the target coal and providing a reliable feature basis for subsequent moisture content calculation.

[0018] Furthermore, the predetermined thermal response characteristics include the time it takes for the surface temperature field of the target coal to reach its peak value after the thermal excitation, the spatial uniformity of the peak temperature distribution, and the decay rate of the temperature field during the cooling phase.

[0019] Preferably, the predetermined thermal response characteristics include the time it takes for the surface temperature field of the target coal to reach its peak after thermal excitation, the spatial uniformity of the peak temperature distribution, and the decay rate of the temperature field during the cooling phase. The time it takes for the surface temperature field to reach its peak is the peak time of the thermal response, which refers to the moment when the surface temperature of the target coal changes from rising to falling after the application of pulsed thermal excitation. This reflects the speed at which the coal absorbs heat and begins to conduct it internally and dissipate it to the environment. Generally, the higher the moisture content, the greater the specific heat capacity, the more pronounced the heating hysteresis effect, and the longer the time required to reach the peak temperature. The spatial uniformity of the peak temperature refers to the time it takes for the surface temperature field to reach its peak after the pulsed thermal excitation. After thermal excitation, the spatial uniformity of the highest temperature values ​​reached at each pixel location on the target coal surface is typically characterized by the ratio of the standard deviation to the mean of the peak temperature matrix. This reflects the uniformity of the moisture content distribution on the coal surface; the more uneven the distribution, the greater the deviation of this uniformity index. The decay rate of the temperature field during the cooling stage refers to the rate at which the surface temperature of the target coal decreases with the natural cooling process after the pulsed thermal excitation. It is taken as the absolute value of the slope of the logarithmic curve of temperature-time during the cooling stage, reflecting the accelerating effect of the latent heat of vaporization of moisture on the cooling process. Generally, the higher the moisture content, the more significant the heat absorption during evaporation, and the faster the decay rate. These multiple dimensions of characteristics collectively characterize the thermal response behavior of coal, providing a multi-dimensional basis for moisture content determination.

[0020] Step S300: The target thermal response characteristic parameters are compared and traversed in the moisture content thermal response characteristic database to obtain the target moisture content.

[0021] In one embodiment, after obtaining the target thermal response characteristic parameters, these parameters are input into a pre-constructed moisture content thermal response characteristic database. The database is then used to iterate through and compare the target parameters with the corresponding thermal response characteristic parameters under different moisture content conditions. By calculating the matching degree between the target characteristic parameters and each thermal response characteristic parameter, feature records whose similarity meets a preset limit are selected, and their corresponding moisture content data is extracted. This moisture content is then used as the target moisture content of the target coal, thereby achieving moisture content identification and quantitative estimation based on thermal response characteristics. Through this method, the thermal response behavior of the target coal under pulsed thermal excitation can be correlated and matched with historically known moisture content samples, thus enabling rapid identification and accurate determination of the surface moisture content of the target coal.

[0022] Furthermore, step S300 is followed by step S400, which includes: Step S410: Iteratively execute steps S100 to S300 to obtain the time series of moisture content of the target coal; Step S420: Perform scatter processing on the time series of moisture content to obtain a scatter plot of moisture content; Step S430: Randomly select the first scatter set in the scatter plot of moisture content and perform regression analysis on the first scatter set to obtain the first regression spline; Step S440: Render the first regression spline onto the scatter plot of moisture content to obtain the first regression plot; Step S450: When the first regression plot reaches a predetermined regression threshold, use the first regression spline as the moisture content curve of the target coal; Step S460: Determine whether the target coal has a predetermined risk based on the moisture content curve, and if so, issue a warning signal.

[0023] Preferably, after obtaining the moisture content result of the target coal under a single test, the infrared continuous monitoring, thermal response feature extraction, and database comparison processes are repeatedly executed at preset time intervals, thereby obtaining the corresponding moisture content at multiple consecutive time points and forming time-series data of the target coal's moisture content based on the detection time sequence. Subsequently, the moisture content detection values ​​corresponding to each time point are correlated with their acquisition time, and discretized with time as the horizontal axis and moisture content as the vertical axis, thereby generating a scatter plot of moisture content reflecting the trend of the target coal's moisture content change. Based on this scatter plot of moisture content, to reduce the interference of random fluctuations, local noise, and abnormal detection points on trend judgment, a portion of the scatter points in the moisture content scatter plot is randomly selected as the first scatter set. This first scatter set is then subjected to regression analysis processing through spline fitting, local weighted regression, piecewise regression, etc., to obtain a first regression spline that can characterize the overall trend of change of this portion of scatter points. Taking spline fitting as an example, the scattered points in the first scatter point set are first sorted according to time order. A discrete data point set is constructed with time as the independent variable and moisture content as the dependent variable. Then, based on the distribution of the scattered points, the positions of several nodes are determined by time intervals, and a piecewise cubic polynomial function is constructed between adjacent nodes, so that each segment of the function fits the scattered points within the corresponding interval. Next, a continuity constraint condition is applied to each segment of the cubic polynomial function, so that the function value, first derivative, and second derivative are continuous at the connection points of adjacent intervals, thereby ensuring the smoothness of the overall fitting curve. At the same time, the coefficients of each segment of the polynomial are solved by the least squares method to minimize the error between the fitted curve and the original scattered points. Finally, the cubic polynomial functions of each segment are spliced ​​to form a continuous regression spline curve as the first regression spline, which is used to characterize the overall trend of the target coal moisture content changing with time.

[0024] After obtaining the first regression spline, it is overlaid on the moisture content scatter plot to form the corresponding first regression plot. The degree of fit and trend representation of the first regression spline to the original scatter plot is then observed. Next, the validity of the first regression plot is judged by a preset regression threshold, determined based on the coverage of the scatter plot by the regression spline. When the first regression plot meets the preset threshold, it indicates that the first regression spline can stably reflect the true trend of the target coal's moisture content over time. In this case, the first regression spline is determined as the moisture content curve of the target coal. Subsequently, based on the variation range of the moisture content curve, a judgment is made on whether the target coal faces a predetermined risk. This predetermined risk includes internal water accumulation, external water seepage, or other safety hazards caused by abnormal changes in moisture content. When the moisture content curve indicates that the target coal faces a predetermined risk, a corresponding early warning signal is output. This early warning signal can be any one or more of the following: audible and visual alarm signals, graphical interface prompts, control terminal alarms, or remote communication alarms, to remind on-site personnel to promptly investigate, handle, or replace the sample for monitoring. By using the above processing method, discrete multiple moisture content detection results can be transformed into a continuous, smooth, and trend-meaning moisture content change curve, thereby improving the stability of coal moisture content anomaly identification and the reliability of early warning judgment.

[0025] Furthermore, step S450 includes: Step S451: Expand the first regression spline by introducing a predetermined distance upper limit to obtain a first regression domain; Step S452: Count the number of scatter points in the first regression domain and record it as the first regression scatter point number; Step S453: Take the ratio of the first regression scatter point number to the total number of scatter points in the moisture content scatter plot as the first regression coefficient; Step S454: When the first regression coefficient is at the predetermined regression threshold, use the first regression spline as the moisture content curve.

[0026] Optionally, after obtaining the first regression spline, to evaluate its coverage and trend representation capabilities of the original scatter points in the moisture content scatter plot, a predetermined upper limit distance is introduced based on the first regression spline, and an envelope is formed by expanding outward from the first regression spline, thus obtaining the first regression domain. This predetermined upper limit distance is used to limit the maximum allowable deviation between the scatter points and the first regression spline. It is set according to the moisture content detection error range, historical experience values, or preset accuracy requirements. All scatter points falling within the first regression domain are considered to have a high degree of consistency with the first regression spline. Subsequently, each scatter point in the moisture content scatter plot is judged point by point, and the number of scatter points located within the first regression domain is counted and recorded as the number of first regression scatter points. Then, the ratio of the number of first regression scatter points to the total number of scatter points in the moisture content scatter plot is calculated to obtain the first regression coefficient. This first regression coefficient is used to characterize the overall fit and coverage ratio of the first regression spline for all moisture content scatter points. Next, the first regression coefficient is compared with a pre-set regression threshold. When the first regression coefficient is not lower than the pre-set threshold, it indicates that the first regression spline can reflect the overall trend of the moisture content scatter plot well within the allowable error range. In this case, the first regression spline is determined as the moisture content curve of the target coal. Otherwise, it is considered that the current first regression spline does not fit the trend of the scatter plot well enough and is not used as the final moisture content curve. Through the above method, the effectiveness of the regression spline can be quantitatively judged by the coverage ratio of the regression domain, thereby improving the objectivity and reliability of the moisture content curve determination process.

[0027] For example, in a single monitoring session, the moisture content of the target coal was measured at 10 consecutive time points, resulting in scatter plot data (unit: %) of moisture content: 6.2, 6.5, 6.7, 6.4, 6.8, 7.0, 6.9, 7.2, 7.1, and 7.3, forming 10 scatter plots. Based on a randomly selected first set of scatter plots, cubic spline fitting was performed to obtain a first regression spline. Using this first regression spline as a benchmark, a predetermined upper limit for distance was set to ±0.3%. Next, the fitted values ​​of the first regression spline at each time point are recorded as: 6.3, 6.4, 6.6, 6.5, 6.9, 6.95, 7.0, 7.15, 7.05, and 7.25. The deviations between each scatter point and the regression spline are calculated point by point, and the deviations are: 0.1, 0.1, 0.1, 0.1, 0.1, 0.05, 0.1, 0.05, 0.05, and 0.05, respectively. All of these deviations are less than the predetermined upper limit of 0.3%. Therefore, all 10 scatter points fall within the first regression domain. At this time, the number of first regression scatter points is 10, and the total number of scatter points is 10. The first regression coefficient is 10 / 10 = 1.0.

[0028] Furthermore, the predetermined risk refers to the presence of internal water accumulation or external water seepage in the coal.

[0029] Optionally, the predetermined risk refers to abnormal moisture content changes in the target coal that do not conform to the normal moisture content during continuous monitoring. This includes two types of risk states: internal water accumulation and external water seepage. Internal water accumulation refers to moisture remaining in the pores, fissures, interlayer structures, or internal areas of the target coal before it fully manifests on the external surface. This causes the target coal to exhibit abnormally delayed temperature rise, lower peak temperature, delayed local cooling, or persistently low temperature characteristics during the thermal response after thermal excitation. External water seepage refers to external moisture seeping into the coal body from the outer surface or local contact areas due to rainfall, spraying, condensation, pipeline leaks, site dampness, or high humidity in the surrounding environment. This results in an abnormally high moisture content state on the surface or in local areas of the target coal. In practical applications, the moisture content can be determined based on the slope of the moisture content curve, the confidence bandwidth of the regression domain, and the scatter point fluctuation noise. When the slope of the moisture content curve slowly drifts positively, the confidence bandwidth of the regression domain widens, and the scatter point fluctuation noise increases, it indicates that the current situation is internal water accumulation. When the slope of the moisture content curve changes sharply by a step, the confidence bandwidth of the regression domain narrows, and the scatter point fluctuation noise decreases, it indicates that the current situation is external water seepage. By defining the predetermined risk as internal water accumulation or external water seepage, the platform can not only achieve numerical detection of the surface moisture content of coal, but also further identify the risk sources causing abnormal moisture content, thereby providing a basis for on-site drainage, sample replacement, pile turning, isolation, or early warning and disposal.

[0030] Furthermore, step S300 includes: Step S310: Extract the first record from the historical coal infrared imaging records, wherein the first record refers to the first experimental dataset obtained by monitoring the moisture content of the target coal sample based on infrared imaging; Step S320: Perform feature traversal on the first experimental dataset based on the predetermined thermal response features to obtain the first thermal response feature parameters; Step S330: Compare the first parameter similarity between the target thermal response feature parameters and the first thermal response feature parameters; Step S340: If the first parameter similarity reaches the predetermined similarity tolerance value, obtain the first moisture content of the target coal sample and use the first moisture content as the target moisture content.

[0031] Preferably, during the moisture content feature matching process, historical coal infrared imaging records are first read from the moisture content thermal response feature database, and a first record matching the current detection conditions is extracted. This first record is a first experimental dataset formed by experimental monitoring of coal samples with known moisture content based on infrared imaging. This first experimental dataset contains the infrared image sequence of the corresponding coal sample under predetermined thermal excitation conditions, environmental parameter information, and the actual moisture content label determined by standard methods. Subsequently, the first experimental dataset is subjected to feature traversal analysis according to the same processing flow as the target coal. That is, based on the predetermined thermal response features, feature information such as the time when the surface temperature field reaches its peak, the spatial uniformity of the peak temperature distribution, and the decay rate during the cooling stage are extracted from the infrared image sequence corresponding to the first experimental dataset. The above features are then processed with unified dimensions and vectorized to obtain the first thermal response feature parameters corresponding to the first record. Afterward, the target thermal response feature parameters and the first thermal response feature parameters are compared item by item, and the first parameter similarity between the two is calculated. This first parameter similarity can be obtained based on weighted Euclidean distance and cosine similarity to reflect the degree of similarity between the target coal and historical coal samples in terms of thermal response behavior. During the calculation process, corresponding weights can be set according to the differences in the sensitivity of different thermal response characteristics to moisture content, thereby improving the discrimination ability of similarity evaluation. When the similarity of the first parameter reaches the preset similarity tolerance value, it indicates that the target coal and the historical coal sample corresponding to the first record have a high consistency in thermal response characteristics. At this time, the known moisture content label corresponding to the first record is obtained and used as the target moisture content of the current target coal. If the similarity of the first parameter does not reach the similarity tolerance value, the next record is extracted from the historical coal infrared imaging record and the above feature extraction and similarity calculation process is repeated until a matching result that meets the conditions is found. Through the above processing method, feature matching-based moisture content determination based on historical samples can be realized, thereby improving the reliability of the detection results and the ability to adapt to different coal types and working conditions.

[0032] Furthermore, the target coal sample and the target coal belong to the same type of coal in the predetermined coal types, which include at least anthracite and bituminous coal.

[0033] Optionally, to improve the accuracy of moisture content identification, when matching historical coal infrared imaging records, a target coal belonging to the same coal type as the current target coal can be selected as a reference object. This ensures that the coal sample used for subsequent feature comparison maintains consistency with the target coal in terms of coal type attributes. This predetermined coal type includes different types such as anthracite and bituminous coal. Since different coal types differ in their microstructure, pore characteristics, and surface thermal response behavior, feature matching within the same coal type can reduce the impact of coal type differences on the moisture content determination results, thereby improving the specificity of thermal response feature comparison and the reliability of moisture content identification.

[0034] Furthermore, step S340 is followed by step S350, which includes: Step S351: Perform enhanced fusion processing on multiple frames of infrared images in the target infrared image sequence to obtain a target fused image; Step S352: Determine whether a predetermined cold spot block exists in the target fused image; Step S353: If it exists, perform feature acquisition and analysis on the target cold spots in the target fused image to obtain a target cold spot coefficient; Step S354: Calibrate the target moisture content using the target cold spot coefficient as a weight; wherein, the predetermined cold spot block refers to the cold spot area generated by the evaporation and cooling effect of the target coal due to uneven local moisture content after the target coal is subjected to the thermal excitation; wherein, the target cold spot coefficient is a coefficient obtained by weighted calculation of the temperature difference and area ratio between the cold spot area and the dry area in the target fused image.

[0035] Preferably, after obtaining the target infrared image sequence, to improve the ability to identify local moisture anomalies, multiple frames of infrared images in the target infrared image sequence are enhanced and fused to generate a target fused image that comprehensively reflects the thermal response characteristics of the target coal. Specifically, a combination of multi-frame superposition averaging and adaptive histogram equalization is used to average the gray values ​​of the same spatial location at different times in the time domain to suppress noise. Then, the contrast of the fused single image is stretched to make the boundary between the low-temperature region formed by the evaporative cooling effect and the surrounding dry and high-temperature region clearer and more distinguishable, thus obtaining a target fused image with a high signal-to-noise ratio. Subsequently, region analysis is performed on the target fused image. By setting temperature thresholds and temperature gradient thresholds, low-temperature regions in the target fused image are detected to determine whether there are predetermined cold point blocks. These predetermined cold point blocks are relatively low-temperature regions in the infrared image caused by the evaporative endothermic effect due to the high local moisture content of the target coal after thermal excitation. If a predetermined cold spot area is detected, feature acquisition and analysis are performed on the cold spot region in the target fused image. This includes extracting parameters such as the temperature difference between the cold spot region and the surrounding dry region, and the area ratio of the cold spot region. These parameters are then normalized and weighted by variation. Specifically, the temperature difference and area ratio after normalization of the maximum and minimum values ​​are used to calculate the first variation weight and the second variation weight. The first variation weight is the normalized result of the temperature difference deviation, and the second variation weight is the normalized result of the area deviation. The target cold spot coefficient is calculated by weighting and summing the temperature difference with the first variation weight and the area ratio with the second variation weight. The temperature difference is the absolute difference between the average gray value of all pixels in the target cold spot region and the average gray value of the dry region in the target fused image, which is then converted into the actual Celsius temperature difference using the calibration curve of the infrared thermal imager. The area ratio is the ratio of the total number of pixels in the target cold spot region to the total number of pixels in the overall effective imaging area of ​​the target coal.

[0036] Subsequently, the target cold point coefficient is incorporated as a weight into the calculation result of the target moisture content, and the original target moisture content obtained through thermal response feature matching is calibrated. That is, a linear weighted correction method is used to ensure that the final output moisture content result can take into account both the overall thermal response characteristics and local abnormal moisture content characteristics. Through the above processing method, cold point phenomena caused by uneven local moisture content can be effectively identified, and the moisture content detection results can be specifically corrected, thereby improving the accuracy of the detection results and the sensitivity to abnormal conditions such as local seepage and water accumulation.

[0037] Furthermore, step S351 is followed by: Determine whether the target fused image has a predetermined surface defect. If it does, perform sample replacement monitoring on the target coal.

[0038] Optionally, after obtaining the target fused image, surface integrity detection is performed on the target fused image to determine whether there are predetermined surface defects on the target coal surface. Specifically, based on the continuity of temperature distribution, texture distribution features, edge morphology features, and local abnormal region morphology in the target fused image, predetermined surface defects are identified on the target coal surface. These predetermined surface defects include abnormal surface states such as surface cracks, holes, spalling, obvious pits, attached foreign objects, and occluded areas. In actual judgment, by extracting the degree of temperature abrupt change, edge density, abnormal low-temperature or high-temperature patch shape features, and regional texture continuity in the target fused image, it is possible to identify whether there are defect areas that do not conform to the normal thermal response pattern of coal surfaces. For example, when the target fused image shows elongated temperature fracture zones, irregularly bounded abnormal patches, or areas with abrupt local thermal response and lack of continuous transition with the surrounding area, it can be determined that there are predetermined surface defects on the target coal surface. Because such surface defects can alter the local heating mode, heat conduction path, and infrared radiation distribution of the target coal, the obtained thermal response characteristics may not accurately reflect the moisture content of the coal itself. Therefore, when a predetermined surface defect is identified, the current test results are not directly used as the final basis for moisture content determination. Instead, a sample replacement monitoring is performed on the target coal. This sample replacement monitoring involves removing the target coal sample with the existing surface defect and selecting a new coal sample with a relatively intact, unobstructed surface that can normally receive thermal excitation and infrared imaging monitoring as the new test object. The aforementioned processes of thermal excitation, infrared image acquisition, thermal response feature extraction, database comparison, and moisture content analysis are then repeated. By adding a surface defect judgment and sample replacement monitoring mechanism, detection bias caused by sample surface abnormalities can be avoided, improving the authenticity, stability, and reliability of moisture content monitoring results.

[0039] In summary, the embodiments of this application have at least the following technical effects: First, an infrared thermal imager is activated to continuously monitor the target coal under thermal excitation, obtaining a sequence of target infrared images. Then, based on predetermined thermal response characteristics, spatiotemporal feature analysis is performed on the target infrared image sequence to obtain target thermal response feature parameters. Finally, the target thermal response feature parameters are compared and analyzed in a moisture content thermal response feature database to obtain the target moisture content. This solves the technical problem of low detection efficiency and difficulty in achieving continuous online monitoring due to existing methods of coal surface moisture content detection relying on contact or manual sampling. It achieves the technical effect of improving the real-time performance, accuracy, and anti-interference capability of coal surface moisture content monitoring by introducing pulsed thermal excitation combined with spatiotemporal feature analysis of infrared image sequences and database comparison.

[0040] Example 2, based on the same inventive concept as the intelligent monitoring method for coal surface moisture content based on infrared imaging in the previous examples, such as... Figure 2As shown, this application provides an intelligent monitoring platform for coal surface moisture content based on infrared imaging. The platform includes: Continuous monitoring module 11: Activates the infrared thermal imager to continuously monitor the target coal under thermal excitation and obtains the target infrared image sequence; Feature analysis module 12: Performs spatiotemporal feature analysis on the target infrared image sequence based on predetermined thermal response features to obtain target thermal response feature parameters; Feature comparison module 13: Traverses and compares the target thermal response feature parameters in the moisture content thermal response feature database to obtain the target moisture content.

[0041] Furthermore, the feature analysis module 12 is used to perform the following methods: The predetermined thermal response characteristics include the time it takes for the surface temperature field of the target coal to reach its peak value after thermal excitation, the spatial uniformity of the peak temperature distribution, and the decay rate of the temperature field during the cooling stage.

[0042] Furthermore, the feature comparison module 13 is used to perform the following method: Step S410: Iteratively execute steps S100 to S300 to obtain the time series of moisture content of the target coal; Step S420: Perform scatter processing on the time series of moisture content to obtain a scatter plot of moisture content; Step S430: Randomly select the first scatter set in the scatter plot of moisture content and perform regression analysis on the first scatter set to obtain the first regression spline; Step S440: Render the first regression spline onto the scatter plot of moisture content to obtain the first regression plot; Step S450: When the first regression plot reaches a predetermined regression threshold, use the first regression spline as the moisture content curve of the target coal; Step S460: Determine whether the target coal has a predetermined risk based on the moisture content curve, and if so, issue a warning signal.

[0043] Furthermore, the feature comparison module 13 is used to perform the following method: Step S451: Expand the first regression spline by introducing a predetermined distance upper limit to obtain a first regression domain; Step S452: Count the number of scatter points in the first regression domain and record it as the first regression scatter point number; Step S453: Take the ratio of the first regression scatter point number to the total number of scatter points in the moisture content scatter plot as the first regression coefficient; Step S454: When the first regression coefficient is at the predetermined regression threshold, use the first regression spline as the moisture content curve.

[0044] Furthermore, the feature comparison module 13 is used to perform the following method: The predetermined risk refers to the presence of internal water accumulation or external water seepage in the coal.

[0045] Furthermore, the feature comparison module 13 is used to perform the following method: Step S310: Extract the first record from the historical coal infrared imaging records, wherein the first record refers to the first experimental dataset obtained by monitoring the moisture content of the target coal sample based on infrared imaging; Step S320: Perform feature traversal on the first experimental dataset based on the predetermined thermal response features to obtain the first thermal response feature parameters; Step S330: Compare the first parameter similarity between the target thermal response feature parameters and the first thermal response feature parameters; Step S340: If the first parameter similarity reaches the predetermined similarity tolerance value, obtain the first moisture content of the target coal sample and use the first moisture content as the target moisture content.

[0046] Furthermore, the feature comparison module 13 is used to perform the following method: The target coal sample and the target coal belong to the same type of coal in the predetermined coal type, which includes at least anthracite and bituminous coal.

[0047] Furthermore, the feature comparison module 13 is used to perform the following method: Step S351: Perform enhanced fusion processing on multiple frames of infrared images in the target infrared image sequence to obtain a target fused image; Step S352: Determine whether a predetermined cold spot block exists in the target fused image; Step S353: If it exists, perform feature acquisition and analysis on the target cold spots in the target fused image to obtain a target cold spot coefficient; Step S354: Calibrate the target moisture content using the target cold spot coefficient as a weight; wherein, the predetermined cold spot block refers to the cold spot area generated by the evaporation and cooling effect of the target coal due to uneven local moisture content after the target coal is subjected to the thermal excitation; wherein, the target cold spot coefficient is a coefficient obtained by weighted calculation of the temperature difference and area ratio between the cold spot area and the dry area in the target fused image.

[0048] Furthermore, the feature comparison module 13 is used to perform the following method: Determine whether the target fused image has a predetermined surface defect. If it does, perform sample replacement monitoring on the target coal.

[0049] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for intelligent monitoring of coal surface moisture content based on infrared imaging, characterized in that, include: Step S100: Activate the infrared thermal imager to continuously monitor the target coal under thermal excitation and obtain the target infrared image sequence; Step S200: Perform spatiotemporal feature analysis on the target infrared image sequence based on predetermined thermal response characteristics to obtain target thermal response feature parameters; Step S300: The target thermal response characteristic parameters are compared and traversed in the moisture content thermal response characteristic database to obtain the target moisture content; The thermal excitation refers to pulsed thermal radiation with a predetermined pulse width and a predetermined pulse intensity. Step S300 is followed by step S400, which includes: Step S410: Iteratively execute steps S100 to S300 to obtain the moisture content time series of the target coal; Step S420: Perform scatter processing on the moisture content time series to obtain a moisture content scatter plot; Step S430: Randomly select the first scatter set in the moisture content scatter plot, and perform regression analysis on the first scatter set to obtain the first regression spline; Step S440: Render the first regression spline onto the moisture content scatter plot to obtain the first regression plot; Step S450: When the first regression plot reaches the predetermined regression threshold, the first regression spline is used as the moisture content curve of the target coal; Step S460: Based on the moisture content curve, determine whether the target coal has a predetermined risk; if so, issue a warning signal. The predetermined thermal response characteristics include the time it takes for the surface temperature field of the target coal to reach its peak value after thermal excitation, the spatial uniformity of the peak temperature distribution, and the decay rate of the temperature field during the cooling stage. Step S300 includes: Step S310: Extract the first record from the historical coal infrared imaging records, wherein the first record refers to the first experimental dataset obtained by monitoring the moisture content of the target coal sample based on infrared imaging; Step S320: Perform feature traversal on the first test dataset based on the predetermined thermal response features to obtain the first thermal response feature parameters; Step S330: Compare the first parameter similarity between the target thermal response feature parameters and the first parameter of the first thermal response feature parameters; Step S340: If the similarity of the first parameter reaches a predetermined similarity tolerance value, then obtain the first moisture content of the target coal sample and use the first moisture content as the target moisture content; Step S340 is followed by step S350, which includes: Step S351: Perform enhancement and fusion processing on multiple frames of infrared images in the target infrared image sequence to obtain a target fused image; Step S352: Determine whether the target fused image contains a predetermined cold spot block; Step S353: If it exists, perform feature acquisition and analysis on the target cold points in the target fusion image to obtain the target cold point coefficient; Step S354: Calibrate the target moisture content using the target cold point coefficient as the weight; The predetermined cold spot block refers to the cold spot area generated by the evaporation and cooling effect of the target coal due to uneven local moisture content after the target coal is subjected to the thermal excitation. The target cold spot coefficient refers to the coefficient obtained by weighted calculation of the temperature difference and area ratio between the cold spot region and the dry region in the target fused image.

2. The intelligent monitoring method for coal surface moisture content based on infrared imaging as described in claim 1, characterized in that, Step S450 includes: Step S451: Expand the first regression spline by introducing a predetermined distance upper limit to obtain the first regression domain; Step S452: Count the number of scatter points in the first regression domain and record it as the number of first regression scatter points; Step S453: Take the ratio of the number of the first regression scatter points to the total number of scatter points in the moisture content scatter plot as the first regression coefficient; Step S454: When the first regression coefficient is at the predetermined regression threshold, the first regression spline is used as the moisture content curve.

3. The intelligent monitoring method for coal surface moisture content based on infrared imaging as described in claim 1, characterized in that, The predetermined risk refers to the presence of internal water accumulation or external water seepage in the coal.

4. The intelligent monitoring method for coal surface moisture content based on infrared imaging as described in claim 1, characterized in that, The target coal sample and the target coal belong to the same type of coal in the predetermined coal type, which includes at least anthracite and bituminous coal.

5. The intelligent monitoring method for coal surface moisture content based on infrared imaging as described in claim 1, characterized in that, Step S351 is followed by: determining whether the target fused image has a predetermined surface defect; if so, performing sample replacement monitoring on the target coal.

6. An intelligent monitoring platform for coal surface moisture content based on infrared imaging, characterized in that, The method for intelligent monitoring of coal surface moisture content based on infrared imaging as described in any one of claims 1-5 includes: Continuous monitoring module: Activates the infrared thermal imager to continuously monitor the target coal under thermal excitation and obtain the target infrared image sequence; Feature analysis module: Performs spatiotemporal feature analysis on the infrared image sequence of the target based on predetermined thermal response features to obtain target thermal response feature parameters; Feature comparison module: It compares the target thermal response feature parameters in the moisture content thermal response feature database to obtain the target moisture content.

Images