Port bulk cargo moisture content monitoring method and system based on spectral analysis

Abstract

The invention relates to the technical field of intelligent management of port bulk cargo storage yards, and discloses a port bulk cargo moisture content monitoring method and system based on spectral analysis, and the method comprises the steps: carrying out the infrared spectrum data collection of the surface of a port bulk cargo stack through an infrared spectrometer carried by an unmanned aerial vehicle, and obtaining the infrared light reflection intensity data; identifying the cargo type of the bulk cargo stack through a spectrum matching algorithm, and selecting a corresponding moisture content calculation model; a plurality of mark points are marked according to the position information of the bulk cargo stack, the unmanned aerial vehicle automatically plans a scanning path and collects infrared light reflection intensity data under a specific wavelength, the selected moisture content calculation model is input, and the moisture content value of the bulk cargo stack is obtained through calculation; and when the water content value is lower than a preset threshold value, a watering instruction is generated and sent to a watering device. According to the scheme, the problems of insufficient precision, poor adaptability, weak linkage and the like of a traditional moisture content monitoring technology can be solved, and a solid technical support is provided for green, safe and efficient operation of a port.

Description

Technical Field

[0001] This invention belongs to the field of intelligent management technology for port bulk cargo yards, and particularly relates to a method and system for monitoring the moisture content of port bulk cargo based on spectral analysis. Background Technology

[0002] Port bulk cargo yards store bulk commodities such as iron ore, coal, grain, and fertilizers. The moisture content of these commodities fluctuates dynamically due to weather changes, storage methods, and cargo characteristics. Changes in moisture content directly affect the accuracy of trade settlements, the stability of cargo quality, loading and unloading safety, and environmental protection. Accurate monitoring of the moisture content of stockpiles plays a crucial role in ensuring port operational safety, reducing trade disputes, maintaining cargo value, and improving operational efficiency; it is an essential component of modern port management.

[0003] Existing monitoring technologies primarily rely on fixed infrared detection equipment, which has limited measurement distance and cannot meet the monitoring needs of tall stacks. Uneven stack surfaces cause fluctuations in measurement distance, leading to decreased data accuracy. Different cargo types exhibit significant differences in material properties, and existing technologies lack intelligent identification and adaptive models for cargo types, making accurate moisture content calculation difficult. Infrared technology can only detect surface moisture and cannot obtain information on the internal moisture distribution of the stack, resulting in incomplete monitoring data. Furthermore, existing systems lack intelligent linkage with dust suppression spraying equipment, making it impossible to accurately adjust dust suppression operations based on real-time moisture content data. The monitoring process is largely manual, resulting in low efficiency and poor data continuity, hindering comprehensive, dynamic, and accurate management of moisture content in the storage yard. These shortcomings severely restrict the efficient operation and safe management of port bulk cargo storage yards, urgently requiring innovative solutions.

[0004] Therefore, there is an urgent need to develop a port bulk cargo moisture content monitoring method and system based on spectral analysis, which can solve the problems of insufficient accuracy, poor adaptability and weak linkage of traditional moisture content monitoring technology, and provide solid technical support for the green, safe and efficient operation of ports. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method and system for monitoring the moisture content of bulk cargo in ports based on spectral analysis. This method can solve the problems of insufficient accuracy, poor adaptability, and weak linkage of traditional moisture content monitoring technologies, providing solid technical support for the green, safe, and efficient operation of ports.

[0006] This invention provides a method for monitoring the moisture content of bulk cargo in ports based on spectral analysis. The method includes the following steps: S1. Use an infrared spectrometer mounted on a drone to collect infrared spectral data on the surface of bulk cargo stacks in the port to obtain infrared light reflection intensity data at different wavelengths. S2. Based on infrared light reflectance intensity data, identify the cargo types of bulk cargo stacks using a spectral matching algorithm; the cargo types include iron ore, coal, grain, and fertilizer. S3. Select the corresponding moisture content calculation model based on the identified cargo type; S4. Based on the location information of the bulk cargo stack, mark the top of the stack and the locations of multiple evenly selected marker points around the stack on the map, and the drone automatically plans the scanning path; S5. The UAV flies along the planned scanning path, adjusts the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack, and collects infrared light reflection intensity data at a specific wavelength. S6. Input the collected infrared light reflection intensity data into the moisture content calculation model selected in S3 to calculate the moisture content value of the bulk cargo stack. S7. When the moisture content is lower than the preset threshold, a water spraying command is generated and sent to the water spraying device. S8. Obtain the environmental parameters of the location of the bulk cargo stack, and calculate the moisture evaporation rate based on the environmental parameters using the Penman formula; S9. Based on the water volume in the water spraying instruction, the water evaporation rate, and the preset threshold, formulate and execute a periodic water spraying plan.

[0007] Furthermore, in S2, based on infrared light reflectance intensity data, a spectral matching algorithm is used to identify the type of goods in the bulk cargo stack, including: The wavelet function approximation method is used to convert discrete spectral data into continuous spectral curves; The similarity between the continuous spectrum and the standard database is calculated using an improved spectral angle matching algorithm.

[0008] Furthermore, the improved spectral angle matching algorithm is calculated using the following formula: ; Where σ represents the spectral angle, i represents the index of the wavelength point, n represents the total number of wavelength points, and λ i S represents the i-th wavelength. c (λ) i S represents the reflectance intensity value of the continuous spectral curve of the cargo to be identified obtained at the i-th wavelength. std (λ) i ) represents the reflection intensity value of the reference spectral curve of a certain type of goods obtained at the i-th wavelength.

[0009] Furthermore, in S3, based on the identified cargo type, the corresponding moisture content calculation model is selected, including: A calculation model based on the multi-band ratio method was selected for iron ore; A calculation model based on the dual-band ratio method was selected for coal. Choose a computational model for grain based on a nonlinear normalization method; A computational model based on principal component analysis was selected for fertilizer.

[0010] Furthermore, in S4, based on the location information of the bulk cargo stack, multiple marker points evenly selected around the top of the stack are marked on the map. The drone automatically plans the scanning path, including: The outline of bulk cargo stacks is identified using the onboard visual sensors of drones; Identify the highest point of the bulk cargo stack based on its outline, and use it as the core marker point; Auxiliary markers are set up at predetermined intervals around the highest point of the bulk cargo stack according to the outline of the bulk cargo stack. Based on the core marker points and auxiliary marker points, the B-spline curve algorithm is used to automatically plan the scanning path.

[0011] Furthermore, in S5, the drone flies along the planned scanning path, and at each marked point, the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack is adjusted to 35cm.

[0012] Furthermore, in S6, the collected infrared light reflection intensity data is input into the moisture content calculation model selected in S3 to calculate the moisture content value of the bulk cargo stack, including: The collected infrared light reflection intensity data is input into the selected moisture content calculation model to obtain the moisture content value of each marked point on the surface of the bulk cargo stack; Based on the moisture content values ​​of each marked point on the surface of the bulk cargo stack, a continuous moisture content distribution model of the entire bulk cargo stack is obtained using the Kriging interpolation method. The moisture content value at any location in the bulk cargo stack is calculated based on the continuous moisture content distribution model.

[0013] This invention also provides a port bulk cargo moisture content monitoring system based on spectral analysis, used to execute the aforementioned port bulk cargo moisture content monitoring method based on spectral analysis. The system includes the following modules: The data acquisition module is used to collect infrared spectral data on the surface of bulk cargo stacks in the port using an infrared spectrometer mounted on a drone, and to obtain infrared light reflection intensity data at different wavelengths. The cargo type identification module is used to identify the cargo type of bulk cargo stacks based on infrared light reflectance intensity data and a spectral matching algorithm; and to select the corresponding moisture content calculation model according to the identified cargo type. The path planning module is used to mark the top of the cargo stack and multiple evenly selected marker points around the cargo stack on the map based on the location information of the bulk cargo stack, and the drone automatically plans the scanning path; The data acquisition module is also used for the UAV to fly along the planned scanning path, adjust the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack, and collect infrared light reflection intensity data at a specific wavelength. The moisture content calculation module is used to input the collected infrared light reflection intensity data into the selected moisture content calculation model to calculate the moisture content value of the bulk cargo stack. The instruction output module is used to generate a watering instruction and send it to the watering device when the moisture content value is lower than a preset threshold.

[0014] The embodiments of the present invention have the following technical effects: This invention utilizes a drone equipped with an infrared spectrometer for non-contact measurement and automatically identifies the type of cargo based on a spectral matching algorithm. A dedicated moisture content calculation model is then selected accordingly. This targeted approach overcomes the interference caused by the different physicochemical properties of various bulk materials (such as iron ore, coal, and grain) on moisture measurement. The basic principle is that different materials have different spectral characteristic signals, and a dedicated model can more accurately resolve the spectral information corresponding to moisture content, thus significantly improving the accuracy and reliability of moisture content monitoring data. Secondly, the drone automatically plans the scanning path and maintains a constant optimal measurement distance, ensuring the standardization and consistency of data collection. Precise control of measurement geometry reduces reflection intensity errors caused by changes in distance and angle. Furthermore, Kriging interpolation is used to spatially interpolate moisture content data at discrete points. By leveraging the spatial correlation of the data, a model reflecting the continuous distribution of moisture content across the entire stockpile is constructed. This not only enables precise control of the overall moisture status of the stockpile but also provides a scientific basis for subsequent differentiated and precise management. When the moisture content at a certain location in the stockpile is detected to be below a preset threshold, the system automatically generates a water spraying command for targeted watering. This effectively suppresses dust while significantly improving water resource utilization efficiency, demonstrating the advantages of intelligent environmental management. Attached Figure Description

[0015] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a port bulk cargo moisture content monitoring method based on spectral analysis provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of a port bulk cargo moisture content monitoring system based on spectral analysis provided in an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0018] This invention provides a method for monitoring the moisture content of bulk cargo in ports based on spectral analysis. Figure 1 This is a flowchart of a port bulk cargo moisture content monitoring method based on spectral analysis provided by an embodiment of the present invention. See also... Figure 1 The method includes the following steps: S1. Use an infrared spectrometer mounted on a drone to collect infrared spectral data on the surface of bulk cargo stacks at the port, and obtain infrared light reflection intensity data at different wavelengths.

[0019] A high-precision infrared spectrometer mounted on a drone is used to collect multidimensional spectral data on the surface of bulk cargo stacks at the port. The drone flies around the cargo stacks, and the spectrometer collects infrared light reflectance intensity data in the wavelength range of 1200-2500 nm at a height of 5-15 meters above the cargo stack surface. During the data collection, the drone dynamically adjusts the measurement distance, and a laser rangefinder can be used to maintain the optimal detection distance to ensure data consistency. Because water molecules have characteristic absorption peaks for near-infrared light (such as the 1450 nm and 1940 nm bands), the reflectance intensity decreases regularly with changes in moisture content. By recording reflectance data at different wavelengths, the system can construct spectral curves of the cargo stack surface.

[0020] S2. Based on infrared light reflection intensity data, identify the type of goods in bulk cargo stacks using a spectral matching algorithm.

[0021] In some embodiments, S2 specifically includes: The wavelet function approximation method is used to convert discrete spectral data into continuous spectral curves; The acquired raw spectral data typically exists as discrete points, meaning that reflection intensity values ​​are recorded only at a few specific wavelength positions. This discreteness can affect subsequent matching accuracy. Therefore, the system uses a wavelet function approximation method to process the discrete spectrum, converting it into a continuous spectral curve. Specifically, let the acquired discrete spectral data be represented as S(λ i ), where λ i This represents the i-th wavelength. A series of wavelet basis functions ψ are introduced. j,k (λ), the original data is expanded into a linear combination of these basis functions, and the calculation formula is as follows: ; Among them, S c (λ) represents the continuous spectral curve of the cargo to be identified, where λ represents the wavelength, j represents the scale parameter, k represents the translation parameter, and ψ j,k (λ) represents the wavelet basis function, c j,k This represents the wavelet basis function ψ at the j-th scale and the k-th translation position. j,k (λ) represents the weighting coefficients. Wavelet functions possess excellent localization properties, effectively smoothing noise while preserving spectral details, thereby enhancing feature extraction capabilities and making subsequent matching more robust.

[0022] The similarity between the continuous spectrum and the standard database is calculated using an improved spectral angle matching algorithm.

[0023] After completing the continuous processing, the system enters the matching stage, the core of which is to calculate the similarity between the spectrum to be identified and various reference spectra in the standard library. The improved spectral angle matching algorithm used is essentially to measure the closeness of two sets of spectral vectors through geometric angles.

[0024] In some embodiments, the improved spectral angle matching algorithm is calculated using the following formula: ; Where σ represents the spectral angle, the smaller the angle, the closer the two are in shape, i.e., the higher the similarity; i represents the index of the wavelength point; n represents the total number of wavelength points; λ i S represents the i-th wavelength. c (λ) i S represents the reflectance intensity value of the continuous spectral curve of the cargo to be identified obtained at the i-th wavelength. std (λ) i σ represents the reflectance intensity value of the reference spectral curve of a certain type of goods obtained at the i-th wavelength. When the calculated σ value is less than a preset threshold, the current sample is determined to have successfully matched the standard spectrum of that type. This method does not rely on changes in absolute reflectance intensity, but focuses on the overall shape of the spectral curve, thus exhibiting strong robustness to fluctuations in illumination conditions and changes in surface roughness.

[0025] The types of goods include iron ore, coal, grain, and fertilizer.

[0026] By converting discrete data into continuous data and combining it with high-dimensional spatial angle calculations, the system can accurately distinguish various complex materials, avoiding misjudgments caused by weak or interfering spectral signals. The resulting automatic identification of cargo types lays the foundation for selecting a suitable moisture content calculation model.

[0027] S3. Select the corresponding moisture content calculation model based on the identified cargo type.

[0028] After identifying the types of bulk cargo, the system needs to select the appropriate moisture content calculation model based on the specific material characteristics to ensure the accuracy of moisture estimation. Different types of bulk cargo exhibit significant differences in physical structure, chemical composition, and spectral response behavior; therefore, a uniform calculation method is insufficient to meet the monitoring needs of multiple cargo types. To address this, this solution designs differentiated modeling strategies for four typical cargo types: iron ore, coal, grain, and fertilizer. By matching their unique spectral response mechanisms, targeted moisture content inversion models are constructed.

[0029] In some embodiments, a calculation model based on the multi-band ratio method is selected for iron ore; For materials like iron ore, which have complex mineral compositions and high surface reflectivity, their spectral signals contain not only moisture information but also characteristic absorption peaks of various iron oxides (such as hematite). Relying solely on a single band or simple ratios is susceptible to interference from mineral background, leading to inaccurate moisture estimation. Therefore, the system employs a calculation model based on multi-band linear regression. This model integrates the reflection intensity and ratio relationships at multiple characteristic wavelengths, as expressed below: ; Among them, M ore The value of R(λ) represents the moisture content of the iron ore, and c0, c1, c2, c3, ... represent the regression coefficients obtained through training with laboratory calibration data. i ) indicates at a specific wavelength λ i The reflection intensity is measured; the number of terms in the formula can be adjusted according to accuracy requirements. This model, by introducing multiple band combinations, captures the absorption characteristics of water molecules in the near-infrared region and uses the characteristic bands of the minerals themselves as a reference to learn and remove background signals, thereby improving the accuracy and robustness of moisture estimation.

[0030] A calculation model based on the dual-band ratio method was selected for coal. Coal is highly hygroscopic, and changes in its moisture content are directly reflected in a decrease in reflectance intensity at specific wavelengths. Water molecules exhibit a strong absorption peak at approximately 1940 nm, and reflectance at this wavelength decreases with increasing moisture content. To eliminate the influence of coal's color, particle size, and surface roughness, the system employs a dual-band ratio method for modeling. The basic form is as follows: ; Among them, M coal Let λ represent the moisture content of the coal, and a and b represent the regression coefficients obtained by measuring a large number of coal samples with known moisture content in the laboratory and using fitting techniques such as the least squares method. R(λ) abs R(λ) is the reflection intensity located near the water absorption peak. refThe selected non-absorption reference band (such as 1300nm or 1800nm) is used for normalization. This ratio can effectively offset reflectance fluctuations caused by non-moisture factors, making the results more focused on changes in moisture content.

[0031] Choose a computational model for grain based on a nonlinear normalization method; Grain grains are typically small, resulting in strong scattering effects when piled up, leading to unstable spectral signals. Traditional linear models struggle to accurately reflect their internal moisture content. Therefore, this solution introduces a nonlinear normalization-based method, employing a logarithmic transformation of the reflection intensity and constructing a normalized difference index to improve stability. Specifically, the normalized difference moisture index is defined as follows: ; Wherein, NDWI represents the Normalized Difference Moisture Index, and ln is the natural logarithm function. Its function is to convert reflectance data to absorbance space, so that the spectral response is closer to the linear relationship of Lambert-Beer law.

[0032] Subsequently, the moisture content M of the grain was determined. grain It is associated with NDWI through a nonlinear function, expressed as follows: ; Where p, q, and r are nonlinear fitting functions, determined through experimental calibration.

[0033] This processing method significantly reduces scattering noise caused by changes in particle morphology and bulk density, enabling the index value to more stably reflect the moisture content inside the grain and enhancing the reliability of the model in complex environments.

[0034] A computational model based on principal component analysis was selected for fertilizer.

[0035] Due to their high hygroscopicity and complex spectral response characteristics, fertilizers often exhibit nonlinear and multivariate coupling features, making them difficult to model effectively using traditional univariate models. Therefore, this system employs Principal Component Analysis (PCA), a data processing method that combines dimensionality reduction and feature extraction. First, all wavelength points in the original spectral data are treated as a high-dimensional vector, and several principal components are extracted using the PCA algorithm: ; Among them, PC m Let w represent the m-th principal component, which is composed of a linear combination of all wavelengths. miThe weight coefficient corresponding to the i-th wavelength in the m-th principal component is automatically calculated by the algorithm, representing the contribution of each wavelength to the principal component. These principal components are arranged in descending order of variance contribution rate. The first few principal components already contain most of the effective information in the spectral data, while redundancy and noise have been removed. Subsequently, these principal components are used as input variables to establish a moisture content prediction model: ; Among them, M fert The values ​​represent the moisture content of the fertilizer, k0, k1, k2, ... k m This represents the regression coefficients obtained through training with laboratory-calibrated data. This method does not directly use the original wavelength data, but instead utilizes "key features" for modeling, making it particularly suitable for handling complex systems that are nonlinear, high-dimensional, and strongly correlated, thus improving the model's generalization ability and robustness.

[0036] In summary, this step, based on the physicochemical properties and spectral behavior of different goods, designed four specialized calculation models: multi-band linear regression, dual-band ratio, nonlinear normalization, and principal component analysis. These models enabled accurate inversion of the moisture content of different types of bulk cargo. Each model is built upon its corresponding physical mechanism and mathematical principles, fully considering interference factors in practical application scenarios, and significantly improving the accuracy and applicability of moisture estimation.

[0037] S4. Based on the location information of the bulk cargo stack, mark the top of the stack and multiple evenly selected marker points around the stack on the map, and the drone automatically plans the scanning path.

[0038] After identifying the cargo types in the bulk cargo stacks, the system needs to further determine the drone's flight and scanning strategies to ensure the comprehensiveness and accuracy of spectral data acquisition. Due to the significant differences in the geometry of different stacks and the complex and variable on-site environment, manual path planning is inefficient and inconsistent, making it difficult to meet the needs of automated monitoring. Therefore, this step integrates visual perception and intelligent path planning technologies to automatically design the drone's scanning path, enabling the flight trajectory to adapt to various stack shapes and avoid potential obstacles.

[0039] In some embodiments, S4 includes the following sub-steps: The outline of bulk cargo stacks is identified using the onboard visual sensors of drones; Identify the highest point of the bulk cargo stack based on its outline, and use it as the core marker point; Auxiliary markers are set up at predetermined intervals around the highest point of the bulk cargo stack according to the outline of the bulk cargo stack. Based on the core marker points and auxiliary marker points, the B-spline curve algorithm is used to automatically plan the scanning path.

[0040] Specifically, firstly, the system uses onboard vision sensors mounted on a drone to acquire images of the bulk cargo stack and identifies its outline boundaries based on computer vision algorithms. This process relies on image edge detection and region segmentation techniques to separate the stack from the background and construct its two-dimensional or three-dimensional geometric model. Based on this, the system uses spatial analysis methods to locate the highest point of the stack, i.e., the top of the cargo pile, as the core monitoring location. This point typically represents a key area for moisture distribution at the top of the stack and is highly representative. Subsequently, multiple auxiliary marker points are evenly distributed around this highest point at preset intervals (e.g., 5 to 10 meters). These points are uniformly distributed on the sides of the stack to cover typical areas of the entire stack surface, thus ensuring the breadth and representativeness of the data collection.

[0041] After obtaining the spatial coordinates of the core marker points and auxiliary marker points, the system uses the B-spline curve algorithm to generate a smooth and continuous flight path, the mathematical expression of which is: ; Among them, P t Let V represent the 3D spatial coordinates (including longitude, latitude, and altitude) of the t-th path point generated by the B-spline curve algorithm. u represents the normalization parameter of the B-spline curve, a continuous variable varying within a specific interval (e.g., [0,1]) used to define continuous positions on the curve. d represents the degree of the B-spline curve, determining its smoothness. S represents the upper limit of the control point index, i.e., the total number of marker points. s represents the index of each control point, with each control point corresponding to one marker point. s B represents the three-dimensional spatial coordinates of the s-th control point. s,d (u) represents the s-th d-th B-spline basis function, determined by the node vectors and the polynomial order, used to control the shape and curvature of the curve. By adjusting the positions of the control points and the parameters of the basis functions, the system can generate a flight trajectory that passes through all key points while maintaining a smooth transition, avoiding measurement instability caused by violent maneuvers.

[0042] S5. The UAV flies along the planned scanning path, adjusts the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack, and collects infrared light reflection intensity data at a specific wavelength.

[0043] Specifically, the drone flies along the planned scanning path, and at each marked point, the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack is adjusted to always remain at 35cm.

[0044] When the UAV flies along the path generated by the B-spline curve to the vicinity of a certain marker point, its onboard visual sensor (or lidar) monitors the relative distance between its current flight altitude and the stack surface in real time. Based on the feedback, the system dynamically adjusts the UAV's heave and descent maneuvers so that when it reaches directly above the marker point, the infrared spectrometer is precisely within the preset measurement distance. Simultaneously, the gimbal performs fine-tuning to ensure that the spectrometer's optical axis is perpendicular to the stack surface, achieving perpendicular incidence and minimizing reflectivity errors caused by oblique incidence. Based on this, the system triggers the spectrometer to collect infrared light reflection intensity data within a specific wavelength range; this data will be used for subsequent moisture content calculations.

[0045] S6. Input the collected infrared light reflection intensity data into the moisture content calculation model selected in S3 to calculate the moisture content value of the bulk cargo stack.

[0046] In some embodiments, S6 includes the following sub-steps: The collected infrared light reflection intensity data is input into the selected moisture content calculation model to obtain the moisture content value of each marked point on the surface of the bulk cargo stack; After acquiring infrared spectral data, the system enters the moisture content inversion and spatial modeling stage. The core task of this step is to transform discrete spectral reflectance intensity data into moisture content information with practical physical meaning, and further construct a continuous model of the moisture distribution on the entire bulk cargo stack surface. Since different cargo types have different spectral response characteristics, the system calls the corresponding moisture content calculation model based on the previous identification results to ensure the scientific rigor and accuracy of the inversion process.

[0047] Based on the moisture content values ​​of each marked point on the surface of the bulk cargo stack, a continuous moisture content distribution model of the entire bulk cargo stack is obtained using the Kriging interpolation method. After obtaining the moisture content values ​​at each marker point, the system faces the challenge of extending these isolated measurements to represent the continuous moisture distribution across the entire stack surface. Since the drone only samples at a limited number of marker points, it cannot cover all areas, and direct interpolation may result in missing or distorted spatial information. Therefore, this solution introduces Kriging interpolation, utilizing the spatial relationships between known points to predict the moisture content at unknown locations. Its mathematical expression is: ; Where M(x,y) represents the estimated moisture content value at the spatial coordinates (x,y) (i.e., the target point where the moisture content needs to be estimated), M s Let w represent the actual measured moisture content value obtained at the s-th known measurement point (i.e., the marker point). s Let M represent the moisture content value at the s-th known measurement point. sThe weight assigned to the measurement point (x, y) when estimating its moisture content is determined by the spatial distance between each measurement point and the target location, as well as the spatial variation structure of the overall data. This weight can be calculated through variogram analysis. Kriging interpolation not only considers the influence of neighboring points but also models the spatial variation trend, thus generating a smoother and more reasonable continuous distribution map.

[0048] The moisture content value at any location in the bulk cargo stack is calculated based on the continuous moisture content distribution model.

[0049] Based on this continuous moisture content distribution model, the moisture content value at any location on the stack can be calculated, realizing the transformation from point to surface. This spatial modeling method significantly improves the visualization and decision support capabilities of monitoring results, enabling managers to intuitively understand the moisture change trend inside the stack, identify high-risk areas (such as localized dryness), and formulate precise watering strategies accordingly.

[0050] S7. When the moisture content is lower than the preset threshold, a water spraying command is generated and sent to the water spraying device.

[0051] When the system determines, using a continuous moisture content distribution model constructed through Kriging interpolation, that the moisture content of a certain area or the entire stack is below a preset threshold, it considers the current state to pose a dust risk and requires the activation of dust suppression measures. The system will automatically generate a water spraying command and send it to a mobile water spraying device deployed near the stack. This command can include parameters such as the location information of the target area, the required spraying intensity, and the duration, ensuring that the water spraying operation can be accurately positioned and executed as needed. After receiving the command, the water spraying device moves to the designated position according to its built-in control system and adjusts the nozzle angle and flow rate to implement directional spraying.

[0052] This invention utilizes a drone equipped with an infrared spectrometer for non-contact measurement and automatically identifies the type of cargo based on a spectral matching algorithm. A dedicated moisture content calculation model is then selected accordingly. This targeted approach overcomes the interference caused by the different physicochemical properties of various bulk materials (such as iron ore, coal, and grain) on moisture measurement. The basic principle is that different materials have different spectral characteristic signals, and a dedicated model can more accurately resolve the spectral information corresponding to moisture content, thus significantly improving the accuracy and reliability of moisture content monitoring data. Secondly, the drone automatically plans the scanning path and maintains a constant optimal measurement distance, ensuring the standardization and consistency of data collection. Precise control of measurement geometry reduces reflection intensity errors caused by changes in distance and angle. Furthermore, Kriging interpolation is used to spatially interpolate moisture content data at discrete points. By leveraging the spatial correlation of the data, a model reflecting the continuous distribution of moisture content across the entire stockpile is constructed. This not only enables precise control of the overall moisture status of the stockpile but also provides a scientific basis for subsequent differentiated and precise management. When the moisture content at a certain location in the stockpile is detected to be below a preset threshold, the system automatically generates a water spraying command for targeted watering. This effectively suppresses dust while significantly improving water resource utilization efficiency, demonstrating the advantages of intelligent environmental management.

[0053] S8. Obtain the environmental parameters of the location of the bulk cargo stack, and calculate the moisture evaporation rate based on the environmental parameters using the Penman formula; S9. Based on the water volume in the water spraying instruction, the water evaporation rate, and the preset threshold, formulate and execute a periodic water spraying plan.

[0054] In one possible implementation, various environmental parameters of the bulk cargo stack location are acquired in real time by integrating a meteorological sensor network or accessing a local port weather station. These parameters include air temperature, relative humidity, wind speed, sunshine duration, and solar radiation. Based on these environmental parameters, an adapted version of the Penman equation is used to calculate the moisture evaporation rate on the stack surface. In this embodiment, parameters are locally corrected to account for the surface characteristics of the port bulk cargo stack, such as the influence of material particle roughness and color on albedo, dynamically integrating the following key factors: Thermal driving factors, based on air temperature and relative humidity, calculate the air's drying capacity, i.e., the saturated vapor pressure difference. The higher the temperature and the lower the humidity, the stronger the air's ability to remove moisture.

[0055] Energy supply factors, based on solar radiation data and the reflectivity of cargo surfaces, estimate the net radiative energy received by the stacked surface, which is the fundamental energy source for moisture evaporation.

[0056] The driving force is wind speed, which directly affects the flow and renewal rate of air on the stack surface. The higher the wind speed, the more conducive it is to water vapor diffusion, thereby accelerating evaporation.

[0057] The above environmental parameters are input into the built-in calculation engine, which calculates the potential moisture evaporation rate on the surface of the bulk cargo stack under the current meteorological conditions in real time by weighted integration of thermal, energy and dynamic effects, in millimeters per day or millimeters per hour.

[0058] In one possible implementation, after obtaining real-time moisture content monitoring data, watering demand commands, and environmental evaporation rates, the system enters the intelligent watering plan formulation stage. The goal of this stage is to achieve on-demand predictive watering, that is, to optimize water resource use while suppressing dust and avoiding ineffective or excessive watering. Specifically, the system constructs a watering decision model that comprehensively considers the following factors: The current moisture content distribution is derived from the Kriging interpolation model of S6; A preset safety threshold for moisture content is set, representing the minimum allowable moisture content for different types of goods, such as 8% for coal and 6% for iron ore. Real-time moisture evaporation rate, calculated from S8; Historical watering effect data records the increase in moisture content and its duration after each watering. Performance parameters of the sprinkler system include flow rate, coverage area, and moving speed.

[0059] When S7 triggers the water spraying command, the system not only performs immediate spraying, but also starts the cycle planning program. Based on the current and future evaporation rates calculated by S8, it predicts the time required for the moisture content of the target area to drop below the safe threshold without intervention. Combining historical data, it simulates the increase in moisture content of the target area and its decay curve over time after performing a standard water spraying operation.

[0060] The next watering time is calculated based on the time required for the moisture content of the target area to drop below the safe threshold and a preset safe buffer time. In this embodiment, the preset safe buffer time is set to 2 hours to ensure early intervention before the moisture content approaches the threshold. The amount of water sprayed per application is dynamically adjusted according to the predicted evaporation rate. During periods of high evaporation, such as midday when temperatures are high and winds are strong, the amount of water sprayed can be increased or the watering interval shortened. During periods of low evaporation, such as at night when temperatures are low and humidity is low, the amount of water sprayed can be reduced or the interval extended.

[0061] For situations where multiple areas of a large storage yard require water simultaneously, the system generates a watering operation path and priority sequence based on the degree to which the moisture content of each area deviates from the threshold, the evaporation rate, and the potential dust risk.

[0062] The system generates a periodic watering plan, including time, location, and water volume suggestions, and sends it to the watering devices, such as intelligent sprinkler trucks or the dispatch center of a fixed sprinkler system. The watering devices execute the plan and report the actual water volume and location to the system after each operation. Based on the comparison with the subsequent monitoring data and predicted values, the system automatically adjusts the evaporation model parameters and the watering effect model, achieving closed-loop adaptive control of planning-execution-feedback-optimization.

[0063] This invention integrates an environmental sensor network to acquire real-time parameters such as temperature, humidity, wind speed, and solar radiation at the location of the stack, and uses an adapted Penman formula model to calculate the moisture evaporation rate. This solves the technical problem of traditional moisture content monitoring systems that only focus on the current moisture state and cannot quantify and predict the natural trend of moisture loss. This technology improves the system's ability to predict future changes in the moisture content of the stack, providing key data support for forward-looking decision-making and reducing the risk of sudden dust storms caused by unexpected rapid decreases in moisture content due to environmental changes. By constructing a dynamic decision-making model based on real-time watering demand, predicted moisture evaporation rate, and safe moisture content thresholds, it generates and executes periodic watering plans. This solves the technical problems of existing watering operations relying on manual experience or simple threshold triggering, resulting in response lags, unscientific watering frequency and volume, insufficient dust suppression sustainability, or water waste. This technology improves the sustainability and stability of dust suppression and moisturizing effects, while significantly improving water resource utilization efficiency and reducing the overall water cost and environmental management burden of port operations; it represents a leap from passive response watering to proactive predictive moisturizing. Through a closed-loop technology system of environmental perception, evaporation prediction, and intelligent planning, the dynamic maintenance of the moisture content of the stacks is achieved, which not only ensures the safety of the port operation environment but also embodies the modern port management concept of green, efficient, and intelligent operation.

[0064] This invention also provides a port bulk cargo moisture content monitoring system based on spectral analysis, used to execute the aforementioned port bulk cargo moisture content monitoring method based on spectral analysis. Figure 2 This is a schematic diagram of a port bulk cargo moisture content monitoring system based on spectral analysis provided in an embodiment of the present invention. (See attached diagram.) Figure 2 The system includes the following modules: The data acquisition module is used to collect infrared spectral data on the surface of bulk cargo stacks in the port using an infrared spectrometer mounted on a drone, and to obtain infrared light reflection intensity data at different wavelengths. The cargo type identification module is used to identify the cargo type of the bulk cargo stack based on the infrared light reflection intensity data and through a spectral matching algorithm; and to select the corresponding moisture content calculation model according to the identified cargo type. The path planning module is used to mark the top of the cargo stack and the positions of multiple evenly selected marker points around the cargo stack on the map based on the location information of the bulk cargo stack, and the drone automatically plans the scanning path; The specific data acquisition module is also used for the UAV to fly along the planned scanning path, adjust the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack, and acquire infrared light reflection intensity data at a specific wavelength; The moisture content calculation module is used to input the collected infrared light reflection intensity data into the selected moisture content calculation model to calculate the moisture content value of the bulk cargo stack. The instruction output module is used to generate a watering instruction and send it to the watering device when the moisture content value is lower than a preset threshold. The evaporation calculation module is used to obtain environmental parameters of the location of the bulk cargo stack, and calculate the moisture evaporation rate based on the environmental parameters using the Penman formula. The cycle setting module is used to formulate and execute a cycle watering plan based on the water volume in the watering instruction, the water evaporation rate, and the preset threshold.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring the moisture content of bulk cargo in ports based on spectral analysis, characterized in that, The method includes the following steps: S1. Use an infrared spectrometer mounted on a drone to collect infrared spectral data on the surface of bulk cargo stacks in the port to obtain infrared light reflection intensity data at different wavelengths. S2. Based on the infrared light reflection intensity data, the cargo type of the bulk cargo stack is identified by a spectral matching algorithm; wherein, the cargo type includes iron ore, coal, grain and fertilizer; S3. Select the corresponding moisture content calculation model based on the identified cargo type; S4. Based on the location information of the bulk cargo stack, mark the top of the stack and the locations of multiple evenly selected marker points around the stack on the map, and the drone automatically plans the scanning path; S5. The UAV flies along the planned scanning path, adjusts the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack, and collects infrared light reflection intensity data at a specific wavelength. S6. Input the collected infrared light reflection intensity data into the moisture content calculation model selected in S3 to calculate the moisture content value of the bulk cargo stack. S7. When the moisture content value is lower than the preset threshold, a watering command is generated and sent to the watering device. S8. Obtain the environmental parameters of the location of the bulk cargo stack, and calculate the moisture evaporation rate based on the environmental parameters using the Penman formula; S9. Based on the water volume in the water spraying instruction, the water evaporation rate, and the preset threshold, formulate and execute a periodic water spraying plan.

2. The method for monitoring the moisture content of bulk cargo in ports based on spectral analysis according to claim 1, characterized in that, In step S2, based on the infrared light reflectance intensity data, the type of goods in the bulk cargo stack is identified using a spectral matching algorithm, including: The wavelet function approximation method is used to convert discrete spectral data into continuous spectral curves; The similarity between the continuous spectrum and the standard database is calculated using an improved spectral angle matching algorithm.

3. The method for monitoring the moisture content of bulk cargo in ports based on spectral analysis according to claim 2, characterized in that, The improved spectral angle matching algorithm is calculated using the following formula: ； Where σ represents the spectral angle, i represents the index of the wavelength point, n represents the total number of wavelength points, and λ i S represents the i-th wavelength. c (λ) i S represents the reflectance intensity value of the continuous spectral curve of the cargo to be identified obtained at the i-th wavelength. std (λ) i ) represents the reflection intensity value of the reference spectral curve of a certain type of goods obtained at the i-th wavelength.

4. The method for monitoring the moisture content of bulk cargo in ports based on spectral analysis according to claim 1, characterized in that, In step S3, based on the identified cargo type, a corresponding moisture content calculation model is selected, including: A calculation model based on the multi-band ratio method was selected for iron ore; A calculation model based on the dual-band ratio method was selected for coal. Choose a computational model for grain based on a nonlinear normalization method; A computational model based on principal component analysis was selected for fertilizer.

5. The method for monitoring the moisture content of bulk cargo in ports based on spectral analysis according to claim 1, characterized in that, In step S4, based on the location information of the bulk cargo stack, the top of the stack and multiple evenly selected marker points around the stack are marked on the map. The drone automatically plans a scanning path, including: The outline of the bulk cargo stack was identified using the onboard visual sensor of the drone; The highest point of the bulk cargo stack is identified based on its outline and used as the core marker point. Auxiliary markers are arranged at predetermined intervals around the highest point of the bulk cargo stack according to the outline of the bulk cargo stack; Based on the core marker points and the auxiliary marker points, the B-spline curve algorithm is used to automatically plan the scanning path.

6. The method for monitoring the moisture content of bulk cargo in ports based on spectral analysis according to claim 5, characterized in that, In step S5, the UAV flies along the planned scanning path, and at each marked point, the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack is adjusted to 35cm.

7. The method for monitoring the moisture content of bulk cargo in ports based on spectral analysis according to claim 1, characterized in that, In step S6, the collected infrared light reflection intensity data is input into the moisture content calculation model selected in step S3 to calculate the moisture content value of the bulk cargo stack, including: The collected infrared light reflection intensity data is input into the selected moisture content calculation model to obtain the moisture content value of each marked point on the surface of the bulk cargo stack. Based on the moisture content values ​​of each marked point on the surface of the bulk cargo stack, a continuous moisture content distribution model of the entire bulk cargo stack is obtained using the Kriging interpolation method. The moisture content value at any location of the bulk cargo stack is calculated based on the continuous moisture content distribution model.

8. A port bulk cargo moisture content monitoring system based on spectral analysis, used to execute the port bulk cargo moisture content monitoring method based on spectral analysis as described in any one of claims 1-7, characterized in that, The system includes the following modules: The data acquisition module is used to collect infrared spectral data on the surface of bulk cargo stacks in the port using an infrared spectrometer mounted on a drone, and to obtain infrared light reflection intensity data at different wavelengths. The cargo type identification module is used to identify the cargo type of the bulk cargo stack based on the infrared light reflection intensity data and through a spectral matching algorithm; wherein the cargo type includes iron ore, coal, grain and fertilizer; The moisture content model selection module is used to select the corresponding moisture content calculation model based on the identified cargo type. The path planning module is used to mark the top of the cargo stack and the positions of multiple evenly selected marker points around the cargo stack on the map based on the location information of the bulk cargo stack, and the drone automatically plans the scanning path; A specific data acquisition module is used for the UAV to fly along the planned scanning path, adjust the measurement distance between the infrared spectrometer and the surface of the bulk cargo stack, and acquire infrared light reflection intensity data at a specific wavelength; The moisture content calculation module is used to input the collected infrared light reflection intensity data into the selected moisture content calculation model to calculate the moisture content value of the bulk cargo stack. The instruction output module is used to generate a watering instruction and send it to the watering device when the moisture content value is lower than a preset threshold. The evaporation calculation module is used to obtain environmental parameters of the location of the bulk cargo stack, and calculate the moisture evaporation rate based on the environmental parameters using the Penman formula. The cycle setting module is used to formulate and execute a cycle watering plan based on the water volume in the watering instruction, the water evaporation rate, and the preset threshold.

Images