Processing characteristic monitoring and analyzing system for early indica rice for rice noodles in storage process

Abstract

The invention relates to the technical field of grain storage monitoring, and particularly discloses a processing characteristic monitoring and analyzing system for early indica rice for rice noodles in a storage process, which comprises the following steps of: generating a three-dimensional quality deterioration risk map according to storage environment data, and performing layered space sampling according to the three-dimensional quality deterioration risk map to obtain a sample set with coordinate identification; performing differential scanning calorimetry analysis on the sample to obtain gelatinization enthalpy value data, and integrating the gelatinization enthalpy value data with space coordinates and risk levels of the sample into a structured database; by taking the space coordinate as a position variable, the gelatinization enthalpy value as a target variable and the risk value as an auxiliary variable, processing by adopting a covariant Kriging algorithm, and generating a continuous aging degree prediction curved surface covering the whole storage space and corresponding uncertainty evaluation data; and finally, constructing a three-dimensional visual distribution map based on a prediction result, calculating a whole warehouse statistical index, identifying a high-confidence local deterioration region, and forming a quality decision report. The problem of monitoring distortion of a traditional method is effectively solved.

Description

A monitoring and analysis system for processing characteristics of early indica rice during storage for rice flour production. Technical Field

[0001] This invention relates to the field of grain storage monitoring technology, specifically to a system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage. Background Technology

[0002] Early indica rice is a key raw material for making rice noodles, and its high amylose content directly affects the final texture of the rice noodles. During storage, environmental factors such as temperature and humidity cause the starch inside the rice grains to retrograde, leading to changes in gelatinization characteristics and consequently reducing the taste, chewiness, and cooking quality of the rice noodles. Therefore, monitoring the processing characteristics of early indica rice during storage is crucial for ensuring the quality of raw materials and the stability of rice noodle production.

[0003] When monitoring stored rice, the sample size required for testing differs significantly from the total amount of material in the warehouse. Furthermore, the uneven distribution of stored rice in the environment creates localized micro-regions with significantly different degrees of starch aging (i.e., "aging hotspots"). Consequently, the trace samples randomly or sparsely extracted from the macroscopic storage space cannot statistically reliably represent the average aging level of the entire batch, nor can they ensure that the most deteriorated local conditions are captured. This "scale mismatch" inherently risks misjudging the overall processing characteristics of the warehouse using traditional point sampling and stored rice measurement methods. This could lead to batches with severe localized deterioration being deemed acceptable, or batches that are generally in good condition with only minor localized changes being prematurely scrapped, resulting in significant errors in storage management decisions and economic losses. Summary of the Invention

[0004] The purpose of this invention is to provide a system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage, in order to solve the problems mentioned above.

[0005] The objective of this invention can be achieved through the following technical solution: a system for monitoring and analyzing the processing characteristics of early indica rice for rice flour storage, comprising: a storage risk map construction module, which acquires the temperature and humidity parameters of multiple sensor nodes deployed in the storage space, and calculates and generates a quality deterioration risk map with three-dimensional coordinate attributes in the storage space by combining the parameter history sequence of each node; an intelligent stratified sampling module, which generates a stratified sampling scheme with spatial coordinate orientation based on the risk level of different coordinate areas in the quality deterioration risk map, and collects a set of rice samples with corresponding spatial coordinates from different risk level areas in the storage space based on the stratified sampling scheme; a thermodynamic feature analysis module, which performs differential scanning calorimetry analysis on each sample in the rice sample set to obtain a thermodynamic dataset characterizing the starch aging state with the gelatinization enthalpy as the core; and a multi-source data fusion and structuring module, which integrates the spatial characteristics of each sample... The system integrates coordinates, risk levels at corresponding locations extracted from the quality deterioration risk map, and gelatinization enthalpy values ​​from the thermodynamic dataset to form a structured spatial sample database containing coordinates, risk values, and thermodynamic values. The spatial interpolation and prediction module uses spatial coordinates from the structured spatial sample database as location variables, gelatinization enthalpy as the target variable, and the corresponding risk value as an auxiliary variable. It employs a covariate kriging algorithm to generate aging degree prediction surface data covering all continuous locations in the storage space, along with corresponding prediction uncertainty assessment data. The visualization analysis and decision report generation module constructs a three-dimensional visualized spatial distribution map of starch aging based on the aging degree prediction surface data. Combining the prediction uncertainty assessment data, it calculates the statistical interval of the overall aging degree, the spatial range of local aging peaks, and the confidence probability from the aging degree prediction surface data, generating a quality decision report.

[0006] As a further aspect of the present invention: the calculation and generation process of the quality deterioration risk map is as follows: based on the geometric structure of the storage space, the geometric structure is discretized into multiple volume units with three-dimensional coordinates, and the parameter data of each sensing node is mapped to the volume unit in which it is located; for each volume unit, the time-weighted average of the historical sequence of temperature and humidity parameters within the volume unit is calculated, and combined with the preset early indica rice starch aging rate response function, the current deterioration driving parameters of the corresponding volume unit are calculated; the three-dimensional coordinates of all volume units in the storage space are integrated with the calculated current deterioration driving parameters to generate a quality deterioration risk map expressed in three-dimensional space and characterized by the value of the driving parameters to represent the level of risk.

[0007] As a further aspect of the present invention, the generation process of the stratified sampling scheme is as follows: the degradation driving parameters of each volumetric unit in the quality degradation risk map are compared with preset high-risk and low-risk thresholds to delineate the coordinate ranges of high-risk, medium-risk, and low-risk regions; different sampling density rules are applied to regions with different risk levels: in high-risk regions, a sampling coordinate is assigned to the center point of each volumetric unit; in medium-risk regions, a sampling coordinate is assigned to the center of the region formed by every four adjacent volumetric units; in low-risk regions, a sampling coordinate is assigned to the center of the region formed by every eight adjacent volumetric units; all assigned sampling coordinates and their source region risk level information are summarized to generate the stratified sampling scheme.

[0008] As a further aspect of the present invention: the collection of rice samples with corresponding spatial coordinates specifically includes: according to each sampling coordinate in the stratified sampling scheme, using a sampling device with spatial positioning function to perform fixed-point sampling at the corresponding three-dimensional position in the storage space to obtain single-point rice samples; during the sampling process, recording the precise three-dimensional coordinates corresponding to each single-point rice sample in real time, and associating them with the risk level determined according to the stratified sampling scheme; integrating all the single-point rice samples collected this time, along with their corresponding three-dimensional coordinates and risk level information, to form a rice sample set.

[0009] As a further aspect of the present invention, the process of acquiring the thermodynamic dataset is as follows: based on the risk level recorded for each sample in the rice sample set, a preset differential scanning calorimetry (DSC) program is dynamically matched, wherein a higher resolution temperature-raising scanning program is used for high-risk samples; after each rice sample undergoes moisture balance pretreatment under the same constant temperature and humidity environment, a preset mass is accurately weighed and packaged in a standard crucible; according to the matched analysis program, differential scanning calorimetry is performed on each packaged sample to obtain the original heat flow curve; for each original heat flow curve, integral calculation is performed under a unified baseline to extract the gelatinization enthalpy value corresponding to each sample, and the coordinates and risk level of the sample to which the gelatinization enthalpy value belongs are identified to form a thermodynamic dataset.

[0010] As a further aspect of the present invention, the formation process of the structured spatial sample database is as follows: for each sample in the rice sample set, a unique identifier including spatial coordinates is created; using the unique identifier as a correlation clue, the risk level matching the coordinates is searched in the quality deterioration risk map, and the gelatinization enthalpy value of the matching identifier is extracted from the thermodynamic dataset; the spatial coordinates, risk level, gelatinization enthalpy value, and unique identifier corresponding to each sample are written into the database table as a complete record; after traversing all samples to complete the record writing, the correlation consistency of all records is verified to generate the final structured spatial sample database.

[0011] As a further aspect of the present invention: the processing using the covariate kriging algorithm specifically includes: calculating the spatial covariant relationship between the enthalpy of gelatinization and the risk value in the entire sample space based on a structured spatial sample database; constructing a kriging interpolation calculation rule with spatial coordinates as the independent variable, the enthalpy of gelatinization as the dependent variable, and the risk value as the spatial drift auxiliary variable based on the spatial covariant relationship; calculating the positions of all unsampled points in the storage space according to the kriging interpolation calculation rule, generating the predicted enthalpy of gelatinization and the corresponding kriging prediction variance for each point; integrating the predicted values ​​of all points into aging degree prediction surface data, and integrating the corresponding prediction variances into prediction uncertainty assessment data.

[0012] As a further aspect of the present invention: the calculation of the spatial covariant relationship between the enthalpy of gelatinization and the risk value in the entire sample space specifically includes: determining multiple different spatial lag distance groups, dividing the spatial coordinates of all samples in the structured spatial sample database into corresponding groups according to their mutual distances; for each spatial lag distance group, calculating the mean square of the difference between the enthalpy of gelatinization values ​​and the mean square of the difference between the risk values ​​of all sample pairs within the group, and obtaining the semivariance of the enthalpy of gelatinization and the semivariance of the risk value corresponding to the group; based on all spatial lag distance groups and the corresponding two sets of semivariance data, respectively fitting and generating a semivariance function of the enthalpy of gelatinization describing the spatial autocorrelation characteristics of the enthalpy of gelatinization, and a semivariance function of the risk value describing the spatial autocorrelation characteristics of the risk value.

[0013] As a further aspect of the present invention: the formation of the quality decision report specifically includes: based on the Kriging prediction variance in the prediction uncertainty assessment data, calculating the confidence interval for the predicted value of each point in the aging degree prediction surface data; identifying and delineating the continuous spatial regions in the aging degree prediction surface data where the predicted value exceeds a preset quality threshold and the lower limit of the confidence interval also exceeds the corresponding threshold as high-confidence local aging peak regions; statistically analyzing the distribution of the predicted values ​​of all points in the aging degree prediction surface data, calculating the predicted values ​​corresponding to the percentiles as the overall warehouse aging degree statistical interval, and simultaneously calculating the total volume and average predicted confidence probability of the high-confidence local aging peak regions; integrating the three-dimensional visualized starch aging spatial distribution map, the spatial coordinate range of the high-confidence local aging peak regions, the overall warehouse aging degree statistical interval, and the total volume and average predicted confidence probability to generate a quality decision report.

[0014] The beneficial effects of the present invention are as follows: (1) The present invention guides intelligent stratified sampling by constructing a three-dimensional quality deterioration risk map, ensuring that the samples can cover the entire storage space with different densities according to the risk level, thus ensuring the spatial representativeness of the samples from the source. Furthermore, the covariate kriging algorithm is used to fuse discrete spatial sampling point data with continuous environmental risk information to generate a continuous prediction surface covering the entire warehouse and quantify the uncertainty of each prediction point. This fundamentally improves the monitoring conclusion from a potentially distorted "point estimate" to a "field assessment" that includes spatial distribution details and statistical confidence. Managers can not only see the overall aging degree, but also accurately locate the scope and severity of "aging hotspots" and know how confident they are in making the judgment, thereby making risk-controllable decisions and avoiding the misuse of raw materials or economic losses caused by misjudgment.

[0015] (2) This invention dynamically generates a risk map by integrating real-time environmental sensor data and historical sequences, enabling the monitoring itself to have risk prediction capabilities. Based on this, the system outputs not a single average value, but a comprehensive decision report including three-dimensional spatial distribution, precise coordinates, volume, and confidence probability of locally exceeding standards areas. This transforms warehouse management from a traditional, extensive model based on overall average values ​​to a precise management model that allows for targeted interventions (such as local ventilation or priority outbound storage) for specific high-risk areas. Simultaneously, this method provides unprecedentedly detailed data support for establishing spatial archives of raw material quality, optimizing inventory rotation strategies, and guiding subsequent differentiated processing, achieving proactive and spatial monitoring and management of early indica rice processing characteristics throughout the entire storage cycle. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 is a system block diagram of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please refer to Figure 1. This invention is a monitoring and analysis system for processing characteristics of early indica rice for rice flour storage, comprising: a storage risk map construction module, which acquires the temperature and humidity parameters of multiple sensor nodes deployed in the storage space, and calculates and generates a quality deterioration risk map with three-dimensional coordinate attributes by combining the parameter history sequence of each node; an intelligent stratified sampling module, which generates a stratified sampling plan with spatial coordinate orientation based on the risk level of different coordinate areas in the quality deterioration risk map, and collects a set of rice samples with corresponding spatial coordinates from different risk level areas in the storage space based on the stratified sampling plan; a thermodynamic feature analysis module, which performs differential scanning calorimetry analysis on each sample in the rice sample set to obtain a thermodynamic dataset characterizing the starch aging state with the gelatinization enthalpy as the core; and a multi-source data fusion and structuring module, which integrates the spatial coordinates of each sample, The risk levels at corresponding locations extracted from the quality deterioration risk map and the gelatinization enthalpy values ​​from the thermodynamic dataset are correlated and integrated to form a structured spatial sample database containing coordinates, risk values, and thermodynamic values. The spatial interpolation and prediction module uses the spatial coordinates in the structured spatial sample database as location variables, the gelatinization enthalpy as the target variable, and the corresponding risk values ​​as auxiliary variables. It employs a covariate kriging algorithm to generate aging degree prediction surface data covering all continuous locations in the storage space, along with corresponding prediction uncertainty assessment data. The visualization analysis and decision report generation module constructs a three-dimensional visualized spatial distribution map of starch aging based on the aging degree prediction surface data. Combining the prediction uncertainty assessment data, it calculates the statistical interval of the overall aging degree of the warehouse, the spatial range of local aging peaks, and the confidence probability from the aging degree prediction surface data, generating a quality decision report.

[0020] In the warehouse risk mapping module, temperature and humidity parameters of multiple sensor nodes deployed within the warehouse space are acquired. Combined with the historical parameter sequences of each node, a quality deterioration risk map with three-dimensional coordinate attributes is calculated and generated within the warehouse space. Specifically, the construction of the quality deterioration risk map begins with the digital representation of the physical warehouse space. First, a corresponding three-dimensional virtual space is created in the computing environment based on the actual warehouse's length, width, and height dimensions. This virtual space is then divided along its length, width, and height at fixed intervals (e.g., 1 meter), discretizing it into a series of regularly arranged volumetric units with unique three-dimensional coordinate indices. Each volumetric unit represents a small area within the warehouse. The physical installation locations of multiple pre-deployed temperature and humidity sensor nodes within the warehouse space are recorded and mapped to the corresponding volumetric unit coordinates in the virtual space. During the monitoring process, these sensor nodes collect temperature (in degrees Celsius) and relative humidity (in percentage) values ​​at their locations at set time intervals (e.g., once per hour), and transmit them to the data center for storage via wired or wireless networks to form a long-term historical parameter sequence for each node.

[0021] For each volumetric unit, the calculation of its degradation driving parameters integrates the current environmental state and historical cumulative effects. The calculation process consists of two steps. First, the time-weighted average of the parameter historical sequence provided by the mapped sensor nodes within the volumetric unit is calculated. Specifically, a weight is assigned to each data point in the historical data, and this weight decreases according to a preset decay rule as the time from the data point to the current moment increases. For example, the rule can be set as follows: the total weight of data within the most recent 24 hours accounts for 70%, the total weight of data from 24 hours ago to 168 hours (one week) accounts for 20%, and the total weight of even earlier data accounts for 10%. This weighting rule is applied to the temperature and humidity sequences respectively to calculate the weighted average temperature and weighted average humidity of the volumetric unit. Second, the calculated weighted average temperature and weighted average humidity are input into a preset early indica rice starch aging rate response function. This function is constructed based on the kinetic principle of starch chemical reactions, and its core logic is: under a set baseline humidity condition, the starch aging rate increases exponentially with increasing temperature; under a set baseline temperature condition, the aging rate increases linearly with increasing humidity. The function is implemented as follows: First, based on the change in weighted average temperature relative to the reference temperature, an exponential function with a natural constant as the base is calculated. Then, based on the change in weighted average humidity relative to the reference humidity, a linear scaling factor is calculated as the humidity influence factor. Finally, the temperature influence factor and the humidity influence factor are multiplied to obtain a comprehensive, dimensionless value, which is the current degradation driving parameter for that volumetric unit. A larger value for this parameter indicates a faster potential rate of starch aging due to environmental conditions at that location.

[0022] Finally, information from all volumetric units is integrated to generate a complete map. Each volumetric unit in the virtual warehouse space is traversed, and its three-dimensional coordinate index (e.g., represented by its ordinal number in the length, width, and height directions) is bound to the degradation driving parameter value calculated in the previous step. The summarization of all volumetric unit "coordinate-driving parameter" pairs constitutes the underlying data for the quality degradation risk map. This data is organized within a three-dimensional spatial framework and can be visualized using computer graphics technology: for example, degradation driving parameters with different numerical ranges are mapped to different colors (e.g., red represents high values, blue represents low values), and the corresponding volumetric units at the corresponding coordinates are colored and rendered in the three-dimensional warehouse model, thereby generating a three-dimensional color map that intuitively displays the relative risk levels at different locations within the warehouse space. The color depth of each point in this map is directly determined by the driving parameter value, realizing the spatial expression of risk information.

[0023] In the intelligent stratified sampling module, a stratified sampling plan with spatial coordinate orientation is generated based on the risk levels of different coordinate regions in the quality degradation risk map. Based on this plan, rice samples with corresponding spatial coordinates are collected from different risk level regions within the storage space. Specifically, the generation of the stratified sampling plan begins with the precise division of spatial risk levels in the quality degradation risk map. The preset high-risk and low-risk thresholds are derived from a large amount of historical experimental data statistically analyzing the correlation between the starch gelatinization characteristics of early indica rice and the subsequent rice flour processing quality (such as toughness and breakage rate). Specifically, by performing regression analysis on the thermal analysis data of rice samples under different storage conditions and the key textural indicators of corresponding rice flour products, the critical range of gelatinization enthalpy that leads to a significant decline in processing quality is determined, and the numerical boundary of the strongly correlated "degradation driving parameter" is calculated in reverse. The degradation driving parameter values ​​of each volumetric unit in the map are compared with two thresholds: if a unit's value is greater than or equal to the high-risk threshold, it is classified as a high-risk unit; if its value is less than the low-risk threshold, it is classified as a low-risk unit; and units with values ​​between the two thresholds are classified as medium-risk units. The coordinate ranges of all units marked with the same risk level are aggregated in three-dimensional space, thus clearly defining the high-risk, medium-risk, and low-risk areas within the storage space.

[0024] After delineating different risk zones, differentiated spatial sampling density rules are applied to assign specific sampling coordinates to each zone. In high-risk zones, to ensure precise capture of potential localized severe deterioration, the sampling rule is set as follows: for each independent volumetric unit within the zone, the three-dimensional coordinates of its geometric center point are calculated, and these coordinates are directly designated as a sampling coordinate. In medium-risk zones, to balance monitoring accuracy and sampling workload, the sampling rule is set as follows: every four adjacent volumetric units (usually forming a larger cuboid) are considered a group, the geometric center point coordinates of this group are calculated, and these coordinates are designated as a sampling coordinate. In low-risk zones, due to the relatively stable and uniform quality status, sampling aims to monitor the baseline level; therefore, a sparser rule is adopted: every eight adjacent volumetric units (forming a larger cube) are considered a group, the geometric center point coordinates of this group are calculated, and these coordinates are designated as a sampling coordinate. After traversing all risk zones according to the above rules and completing the calculation and allocation of all sampling coordinates, a list is generated. The list details the three-dimensional coordinates of each planned sampling point and clearly indicates its risk level. This list constitutes the final stratified sampling plan used to guide on-site sampling.

[0025] Based on the generated stratified sampling plan, spatially directional sample collection is conducted at the storage site. Collection is performed using a deep sampling device with spatial positioning capabilities. This device typically integrates a rangefinder and angle encoder. When its probe is lowered from a pre-set inlet at the top of the storage silo to a designated depth, it can calculate and record the three-dimensional coordinates of the probe tip (i.e., the sampling point) within the silo in real time, based on the lowering length, inclination angle, and known coordinates of the inlet at the top of the silo. Operators, following the coordinates on the sampling plan list, operate the device to reach each designated three-dimensional location for pinpoint sampling, obtaining a representative rice sample from that specific location—a single-point rice sample. Upon completion of each sampling action, the collection system automatically records and saves the precise three-dimensional coordinate data measured by the device at that moment, and simultaneously retrieves the pre-set risk level information for that coordinate point from the sampling plan, binding it to the sample.

[0026] To ensure precise sampling based on the preset three-dimensional coordinates (X, Y, Z) in the stratified sampling scheme, the "deep sampling device with spatial positioning function" is not a simple grain sampler, but an integrated system. Specifically, the device consists of a multi-section, telescopic, or rotatable hollow sampling probe, a three-dimensional positioning platform (including precision slide rails or rotating arms along the horizontal X and Y axes, and a depth encoder controlling the vertical lowering of the probe along the Z axis) mounted on a base with preset sampling holes on the silo top, and a central controller. During sampling, the central controller reads the target coordinates from the stratified sampling scheme and automatically drives the three-dimensional positioning platform to move the probe tip to the corresponding horizontal coordinate position within the silo. Subsequently, the depth encoder monitors the probe's lowering depth in real time. When the depth value matches the vertical value of the target coordinates, the sampling head at the bottom of the probe automatically opens, intercepting a rice sample at the precise coordinate point. Throughout the process, the device records the spatial coordinates of the probe tip in real time and binds them to the sample, thereby ensuring that the physical location of each sample strictly corresponds to the theoretical coordinates in the sampling plan, realizing a closed loop of spatial coordinates from the risk map to the physical sample.

[0027] After sampling at all planned coordinate points, all acquired single-point rice samples were collected and organized. Each sample was assigned a unique number, and a correspondence was established between it and the coordinate data recorded during collection and the associated risk level information. All this information, including the physical sample, sample number, corresponding three-dimensional coordinates, and risk level, was systematically integrated into a structured set. This set not only includes physical samples collected from different risk levels within the storage space but also the spatial "identity card" information for each sample, namely its origin coordinates and risk context. This set is defined as the "rice sample set" used for subsequent laboratory analysis. This process ensures that laboratory testing data can be accurately traced back to the specific location and risk context within the storage space.

[0028] In the thermodynamic feature analysis module, differential scanning calorimetry (DSC) is performed on each sample in the rice sample set to obtain a thermodynamic dataset characterizing the starch aging state, with the gelatinization enthalpy as the core. Specifically, the acquisition of the thermodynamic dataset begins with matching a targeted analysis program to each sample in the rice sample set. Two DSC programs are preset for different risk levels of the sample records. The main difference lies in the rate of temperature rise and the data acquisition density. For samples marked as high-risk, due to the potential for more complex starch aging states or more subtle thermal transition processes, the analysis program is set to use a slower heating rate (e.g., 5 degrees Celsius per minute) and a higher data acquisition frequency of the heat flow signal (e.g., 10 data points per second) within the critical temperature range for starch gelatinization (e.g., from 50°C to 100°C) to obtain a higher resolution raw heat flow curve. For samples of medium and low risk levels, a standard temperature rise procedure is used (e.g., 10 degrees Celsius per minute, 5 data points per second). Based on the risk level associated with each sample number in the sample set list, the corresponding analysis procedure parameters are automatically matched.

[0029] Before analysis, all rice samples must undergo standardized pretreatment to eliminate the interference of initial moisture differences on the thermal analysis results. Each sample is placed under constant environmental conditions (e.g., in a desiccator at 25°C and 60% relative humidity) for at least 48 hours to allow its moisture to reach equilibrium. After pretreatment, a predetermined mass of powder (e.g., 5.0 mg, accurate to 0.1 mg) is accurately weighed from each moisture-equilibrated sample using a precision balance. The weighed powder sample is then placed tightly and evenly in a dedicated standard aluminum crucible and sealed using a tableting tool and crucible lid to ensure isolation from the external environment during testing.

[0030] Subsequently, differential scanning calorimetry (DSC) is performed according to the analytical program matched for each sample in the previous step. The sealed sample crucible and an empty, reference crucible of the same type are placed at the sample end and reference end of the DSC, respectively. Under a specified atmosphere (usually high-purity nitrogen), the sample and reference are heated synchronously according to the programmed heating rate. The instrument monitors and records in real time the compensating heat flux power required to maintain the same temperature between the sample and reference ends; the continuous curve of this power changing with time (or temperature) is the original heat flux curve. This curve fully records the endothermic or exothermic effects of physical or chemical changes (such as starch gelatinization) in the sample throughout the entire heating process.

[0031] After obtaining the original heat flow curves, the core thermodynamic parameters, specifically the enthalpy of gelatinization, are extracted through data processing. First, the onset and termination temperatures of the starch gelatinization endothermic peak on each curve are determined. Then, a straight line between the onset and termination temperatures is used as a baseline on the heat flow curve, representing the theoretical heat flow trajectory assuming no thermal effect occurs. Next, the area of ​​the closed region enclosed by the endothermic peak curve and the baseline is calculated. This area is calculated using a numerical integration method: for all data points collected within the onset to termination temperature range, the difference between the heat flow value at each data point and the corresponding temperature point on the baseline is calculated, multiplied by the time interval represented by that data point, and finally, the products of the differences and time at all points are summed to obtain the peak area. This peak area value, after conversion using the instrument's thermal calibration coefficient, yields the enthalpy of gelatinization expressed in units of heat (e.g., joules per gram). Finally, the calculated enthalpy of gelatinization is linked to the sample number that generated the curve, the corresponding spatial coordinates of the sample, and the risk level information to form a complete record. By iterating through all samples in the sample set and summarizing all the records, the final thermodynamic dataset is formed.

[0032] In the multi-source data fusion and structuring module, the spatial coordinates of each sample, the risk level of the corresponding location extracted from the quality deterioration risk map, and the gelatinization enthalpy value from the thermodynamic dataset are correlated and integrated to form a structured spatial sample database containing coordinates, risk values, and thermodynamic values. Specifically, the formation of the structured spatial sample database begins with the precise correlation and integration of multi-source data. First, a unique identifier is created for each sample from the "rice sample set". The method for constructing this identifier is as follows: extract the three-dimensional spatial coordinate values ​​stored in the sample record. Each coordinate value consists of its ordinal number in the length, width, and height directions of the storage area. Arrange these three ordinal numbers in a fixed order (e.g., length or width first, then height), and insert a specific separator (e.g., an underscore "_") between adjacent ordinal numbers to combine these numbers into a unique string. For example, a sample located at the center of the 5th volume unit in the length direction, the 12th in the width direction, and the 3rd in the height direction will have the unique identifier "5_12_3". This identifier will serve as the core index for this sample in all subsequent data processing.

[0033] Subsequently, using the "unique identifier" generated in the previous step as a correlation clue, corresponding attribute information is extracted from the "quality degradation risk map" and the "thermodynamic dataset," respectively. Specifically, the process is as follows: First, the identifier string is parsed and converted back into three-dimensional coordinate values. Next, the data structure of the quality degradation risk map is searched based on these three-dimensional coordinate values. The quality degradation risk map is stored in the computer as a three-dimensional array, where each index corresponds to a three-dimensional sequence number of a volume unit, and the value stored in the array is the "degradation driving parameter" for that unit. By directly indexing the corresponding element in the array using the coordinate value, the degradation driving parameter value at that location can be obtained. Then, according to the preset risk level classification rules (i.e., the comparison with high-risk and low-risk thresholds), the corresponding "risk level" (high-risk, medium-risk, or low-risk) is determined by the driving parameter value. Simultaneously, in the list or table of the thermodynamic dataset, the data row whose "sample number" or internal correlation field matches the current "unique identifier" is searched, and the calculated "gelatinization enthalpy" is extracted from that row.

[0034] Then, all successfully associated data items are organized into a structured record and written to a unified database table. This database table has several predefined fields, primarily including: "Unique Identifier," "Spatial Coordinates_X" (length direction), "Spatial Coordinates_Y" (width direction), "Spatial Coordinates_Z" (height direction), "Deterioration Driving Parameter Value," "Risk Level," and "Enthalpy of Gelatinization." For each sample processed, a new record is inserted into this table, with all fields filled with the specific values ​​or category labels obtained in the previous steps. For example, for a sample with the unique identifier "5_12_3," its record might contain: Spatial Coordinates_X = 5, Spatial Coordinates_Y = 12, Spatial Coordinates_Z = 3, a specific number for the degradation driving parameter (e.g., 15.7), a risk level of "Medium Risk," and another specific number for the enthalpy of gelatinization (e.g., 8.2 joules / gram).

[0035] Finally, after all sample records are written to the database table, a comprehensive consistency check is performed to generate the final reliable structured spatial sample database. The check mainly includes two aspects: first, a "coordinate consistency" check, which verifies whether the coordinates parsed from the "unique identifier" in each record are completely consistent with the values ​​of the "spatial coordinates_X", "spatial coordinates_Y", and "spatial coordinates_Z" fields in the same record, and whether these coordinates are within a reasonable range of the storage space. Second, a "logical consistency" check, which verifies whether the "risk level" in each record conforms to the preset threshold classification rules with its corresponding "deterioration driving parameter value" (for example, if the parameter value is 15.7, and the low-risk threshold is 10 and the high-risk threshold is 20, then the risk level corresponding to this parameter value should be "medium risk"). After all checks pass, this integrated table containing the spatial location information, environmental risk information, and laboratory thermodynamic information of all valid sample points is confirmed as the final "structured spatial sample database," providing a complete and consistent data foundation for subsequent spatial statistical analysis.

[0036] In the spatial interpolation and prediction module, spatial coordinates from the structured spatial sample database are used as location variables, enthalpy of aging as the target variable, and the corresponding risk value as an auxiliary variable. A covariate kriging algorithm is employed to generate aging prediction surface data covering all continuous locations within the storage space, along with corresponding prediction uncertainty assessment data. Specifically, based on the established "structured spatial sample database," a covariate kriging algorithm is used. The core of this approach is to construct a spatial interpolation framework that comprehensively utilizes the spatial location of sampling points, measured enthalpy of aging, and the risk value (i.e., the degradation driving parameter) as an auxiliary variable to predict the enthalpy of aging at any location within the entire storage space and assess the uncertainty of the prediction. The entire process first requires quantifying the spatial structural relationship between the research variable (enthalpy of aging) and the auxiliary variable (risk value).

[0037] The first step is to calculate the spatial covariance between the enthalpy of gelatinization and the risk value across the entire sample space. This relationship is characterized by establishing a "semivariance function" for each variable, which describes how the spatial autocorrelation of the variables changes with increasing distance between sampling points. In practice, a series of increasing spatial lag distance groups are first determined. For example, the first distance group is set to 0 to 2 meters, the second to 2 to 4 meters, and so on. The spatial coordinates of all samples in the database are paired, the straight-line distance between each pair of samples is calculated, and the samples are assigned to the corresponding lag distance group based on this distance. Then, for each group, the experimental semivariance of the enthalpy of gelatinization and the risk value is calculated. For the enthalpy of gelatinization, the formula is: calculate the square of the difference in enthalpy of gelatinization between each pair of samples within the group, then calculate the average of all these squared values, and finally divide this average by 2. This can be expressed mathematically as: ;in, Indicates the lag distance The semivariance of the gelatinization enthalpy experiment. It is the total number of valid sample pairs within the lag distance group. and They are a pair of objects approximately [distance missing] Sample points and Measured value of gelatinization enthalpy. Summation symbol. This indicates all groups within that group. The samples are summed. The same calculation process is performed on the risk values ​​to obtain the corresponding experimental semivariance value sequence. Finally, curve fitting was performed on the relationship between the experimental semivariance values ​​of the enthalpy of gelatinization and the risk value and the lag distance, respectively. Theoretical models such as the spherical model or the exponential model were used for fitting to obtain continuous functions describing their spatial autocorrelation structure, namely the theoretical semivariance function of the enthalpy of gelatinization and the theoretical semivariance function of the risk value.

[0038] The second step is to construct a specific covariate kriging interpolation rule based on the aforementioned spatial covariate function relationship. This rule aims to predict an unsampled point. The core idea is that the predicted value of a point is obtained by linearly weighting the enthalpy of ... Meanwhile, the solution process will directly provide the minimum prediction error variance corresponding to the prediction, i.e., the Kriging variance, which is a measure of prediction uncertainty.

[0039] The third step is to apply the interpolation calculation rules established in the previous step to traverse and calculate the locations of all unsampled grid points within the storage space. First, based on the spatial coordinates of each unsampled point, the risk value corresponding to that point is extracted from the quality degradation risk map. Then, based on the spatial distance between that point and all known sampled points, the Kriging interpolation rules established in the second step are called (i.e., using the already solved weight calculation logic, but requiring the reselection of neighboring points and possible recalculation of local weights for each new point) to calculate the predicted value of the gelatinization enthalpy at that point. This calculation process involves finding several known sampled points closest to the point to be predicted, obtaining the measured gelatinization enthalpy, risk value, and spatial coordinates of these points, and then using the constructed covariance Kriging equation system (which includes the covariance matrix calculated from the semivariance function) to solve for the optimal weights corresponding to the current point to be predicted for this set of sampled points. Finally, these weights are multiplied by the measured gelatinization enthalpy of the corresponding sampled points and summed to obtain the predicted value. Simultaneously, using the same equation system, the Kriging variance of the predicted value at that point can be obtained. The magnitude of the Kriging variance reflects the reliability of the prediction. The larger the value, the higher the uncertainty of the prediction at that location due to the distance from the known sampling point or the complexity of the data spatial structure.

[0040] The fourth step is to integrate the calculation results from all points to form the final data product. The predicted enthalpy of gelatinization for each grid point within the storage space, calculated in the third step, is arranged and organized according to its three-dimensional coordinates, forming a continuous data array that completely covers the three-dimensional space of the storage area. This data array is the "aging degree prediction surface data." It intuitively displays the spatial distribution of starch aging degree in numerical form, with areas of high values ​​corresponding to areas of severe aging. Simultaneously, the kriging variance values ​​corresponding to each grid point are also arranged and organized according to the same coordinate order, forming another data array, namely the "prediction uncertainty assessment data." This data array corresponds to the prediction surface data, indicating the reliability of each value in the prediction surface. These two datasets together constitute the optimal spatial estimate of the spatial heterogeneity of the entire stored rice processing characteristics (using enthalpy of gelatinization as a proxy indicator) and the degree of cognitive understanding based on the current sampled data and environmental risk information.

[0041] In the visualization analysis and decision report generation module, a three-dimensional visualized spatial distribution map of starch aging is constructed based on the aging degree prediction surface data. Combined with prediction uncertainty assessment data, the statistical interval of the overall aging degree, the spatial range of local aging peaks, and the confidence probability are calculated from the aging degree prediction surface data to generate a quality decision report. Specifically, the quality decision report is the final step in the monitoring and analysis process, its core being the transformation of spatial prediction data into structured information usable for warehouse management and processing decisions. The first step in report generation is to calculate a confidence interval for each predicted value in the aging degree prediction surface data based on the prediction uncertainty assessment data. The confidence interval characterizes the reliable range of the predicted value. For any grid point in the surface data, its gelatinization enthalpy prediction value is a specific numerical value, and the prediction uncertainty assessment data provides the Kriging prediction variance for that point. The standard error of the predicted value at that point is calculated, which is the square root of the Kriging prediction variance. Then, at a given confidence level (e.g., 95%), this standard error is multiplied by a standard normal distribution quantile multiplier corresponding to the preset confidence level (e.g., approximately 1.96 for a 95% confidence level) to obtain an error range. Finally, this error range is subtracted from the predicted value at that point to obtain the lower bound of the confidence interval; the error range is added to the predicted value to obtain the upper bound of the confidence interval. This calculation is performed on all points in the surface data, thereby attaching a reliability measure expressed in interval form to the predicted value for each spatial location.

[0042] The second step is to identify and delineate high-confidence localized aging peak areas within the storage space that require special attention. This step requires setting a "quality threshold," derived from the upper limit requirement of the raw material gelatinization enthalpy for rice noodle processing, determined through prior process experiments, for example, set at 8.5 joules per gram. The identification process involves traversing and filtering all grid points: first, checking whether the predicted gelatinization enthalpy of the point exceeds the preset quality threshold; second, and more importantly, checking whether the lower limit of the confidence interval of the predicted value calculated in the first step also exceeds the quality threshold. Only points that simultaneously meet both conditions are considered high-confidence "exceeding" points. Then, spatial clustering analysis is performed on all points marked as high-confidence exceeding the threshold in three-dimensional space, merging spatially adjacent points (e.g., sharing faces, edges, or corners) to form independent continuous spatial regions. Each such continuous region is defined as a "high-confidence localized aging peak area," meaning there is sufficient confidence that the raw materials within this region are no longer suitable for the production of the target rice noodle product.

[0043] The third step involves calculating a series of quantitative statistical indicators reflecting both overall and local conditions. First, for the entire storage space, the frequency distribution of predicted values ​​for all grid points in the aging degree prediction surface data is statistically analyzed. Based on this distribution, the predicted value corresponding to a specific percentile is calculated. For example, the predicted value corresponding to the 95th percentile is calculated and used as the upper limit of the "statistical interval for overall warehouse aging degree"; the 50th percentile (i.e., the median) is calculated as the central trend value. These two values ​​together describe the main distribution range of the overall warehouse material aging degree. Second, spatial measurement calculations are performed for all "high-confidence local aging peak areas" identified in the previous step. The volume of each area is calculated by counting the number of grid points constituting the area and multiplying it by the actual storage space volume represented by each grid point (e.g., each grid represents 1 cubic meter). The volumes of all such areas are summed to obtain the "total volume." Simultaneously, the average predicted confidence probability (derived from Kriging variance and confidence level) of all grid points within these areas is calculated to obtain the "average predicted confidence probability," used to quantify the overall grasp of the judgment regarding these local deterioration areas.

[0044] Finally, all the above analysis results are systematically integrated to generate the final quality decision report. The report first includes a 3D visualization of the spatial distribution of starch aging based on aging degree prediction surface data. This map visually displays the spatial differences in aging degree using color gradients. The report lists the spatial coordinate range of each identified "high-confidence local aging peak area" in text and tables, for example, "located in the northeast corner of the warehouse, with coordinates ranging from (X1, Y1, Z1) to (X2, Y2, Z2)". The report explicitly provides the calculated "statistical interval of aging degree for the entire warehouse" (e.g., median value of 7.2 joules per gram, 95th percentile of 9.8 joules per gram), as well as the "total volume" (e.g., total of 25 cubic meters) and "average prediction confidence probability" (e.g., 92%) of the local peak areas. This information together constitutes a comprehensive assessment conclusion, which can clearly indicate the overall quality level of stored materials, accurately pinpoint the location and scale of local problem areas, and quantify the statistical reliability of the relevant conclusions, thereby providing a direct and reliable basis for subsequent classification, batch processing, or remedial measures.

[0045] The working principle of this invention is as follows: First, the warehouse risk map construction module calculates and generates a quality deterioration risk map with three-dimensional coordinate attributes based on the temperature and humidity parameters and their historical sequences collected by sensor nodes deployed within the warehouse space. Then, the intelligent stratified sampling module generates and executes a spatially oriented stratified sampling plan according to the risk levels of different areas in the map, collecting rice sample sets with precise coordinate labels from areas of different risk levels. The thermodynamic feature analysis module performs differential scanning calorimetry analysis on each sample to obtain a thermodynamic dataset with the enthalpy of gelatinization as the core. The multi-source data fusion and structuring module then integrates and correlates the spatial coordinates, corresponding risk level, and enthalpy of gelatinization of each sample to form a structured spatial sample database. Based on this, the spatial interpolation and prediction module uses the coordinates in this database as the location variable, the enthalpy of gelatinization as the target variable, and the risk value as the auxiliary variable, employing a covariate kriging algorithm to generate aging degree prediction surface data covering the entire warehouse space and corresponding prediction uncertainty assessment data. Finally, the visualization analysis and decision report generation module constructs a three-dimensional visualized spatial distribution map of starch aging based on the predicted data, and calculates the overall warehouse statistical interval by combining uncertainty assessment data, identifies high-confidence local aging peak areas and their confidence probabilities, and generates a comprehensive quality decision report, thereby achieving accurate monitoring and risk assessment of the spatial heterogeneity of the processing characteristics of stored rice.

[0046] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A system for monitoring and analyzing the processing characteristics of early indica rice during storage for rice flour production, characterized in that, include: The warehouse risk map construction module obtains the temperature and humidity parameters of multiple sensor nodes deployed in the warehouse space, and calculates and generates a quality deterioration risk map with three-dimensional coordinate attributes in the warehouse space by combining the parameter history sequence of each node. The intelligent stratified sampling module generates a stratified sampling plan with spatial coordinate orientation based on the risk level of different coordinate areas in the quality deterioration risk map. Based on the stratified sampling plan, a set of rice samples with corresponding spatial coordinates is collected from different risk level areas in the storage space. The thermodynamic feature analysis module performs differential scanning calorimetry (DSC) analysis on each sample in the rice sample set to obtain a thermodynamic dataset characterizing starch aging state, with the gelatinization enthalpy as the core. The multi-source data fusion and structuring module integrates the spatial coordinates of each sample, the risk level of the corresponding location extracted from the quality deterioration risk map, and the gelatinization enthalpy from the thermodynamic dataset to form a structured spatial sample database containing coordinates, risk values, and thermodynamic values. The spatial interpolation and prediction module uses the spatial coordinates in the structured spatial sample database as the location variable, the gelatinization enthalpy as the target variable, and the corresponding risk value as an auxiliary variable, employing a covariate kriging algorithm to generate aging degree prediction surface data covering all continuous locations in the storage space, along with corresponding prediction uncertainty assessment data. The visualization analysis and decision report generation module constructs a three-dimensional visualized spatial distribution map of starch aging based on the aging degree prediction surface data. Combining the prediction uncertainty assessment data, it calculates the statistical interval of the overall aging degree in the storage area, the spatial range of local aging peaks, and the confidence probability from the aging degree prediction surface data, generating a quality decision report.

2. The system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage according to claim 1, characterized in that, The calculation and generation process of the quality deterioration risk map is as follows: Based on the geometric structure of the storage space, the geometric structure is discretized into multiple volume units with three-dimensional coordinates, and the parameter data of each sensing node is mapped to the volume unit in which it is located; for each volume unit, the time-weighted average of the historical sequence of temperature and humidity parameters in the volume unit is calculated, and combined with the preset early indica rice starch aging rate response function, the current deterioration driving parameters of the corresponding volume unit are calculated; The three-dimensional coordinates of all volume units within the storage space are integrated with the calculated current degradation driving parameters to generate a quality degradation risk map that is expressed in three-dimensional space and uses the values ​​of the driving parameters to characterize the level of risk.

3. The system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage according to claim 1, characterized in that, The generation process of the stratified sampling scheme is as follows: The degradation driving parameters of each volumetric unit in the quality degradation risk map are compared with preset high-risk and low-risk thresholds to delineate the coordinate ranges of high-risk, medium-risk, and low-risk areas; different sampling density rules are applied to areas with different risk levels: in high-risk areas, a sampling coordinate is assigned to the center point of each volumetric unit; in medium-risk areas, a sampling coordinate is assigned to the center of the area formed by every four adjacent volumetric units; in low-risk areas, a sampling coordinate is assigned to the center of the area formed by every eight adjacent volumetric units; all assigned sampling coordinates and their source area risk level information are summarized to generate the stratified sampling scheme.

4. The system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage according to claim 1, characterized in that, The collection of rice samples with corresponding spatial coordinates specifically includes: using a sampling device with spatial positioning function to perform fixed-point sampling at the corresponding three-dimensional position within the storage space according to each sampling coordinate in the stratified sampling plan, and obtaining single-point rice samples; during the sampling process, recording the precise three-dimensional coordinates corresponding to each single-point rice sample in real time, and recording the risk level determined according to the stratified sampling plan; and integrating all the single-point rice samples collected this time, along with their corresponding three-dimensional coordinates and risk level information, to form a rice sample set.

5. The system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage according to claim 1, characterized in that, The process of acquiring the thermodynamic dataset is as follows: based on the risk level recorded for each sample in the rice sample set, a preset differential scanning calorimetry program is dynamically matched, wherein a higher resolution temperature-raising scanning program is used for high-risk samples; after each rice sample is pretreated for moisture balance under the same constant temperature and humidity environment, a preset mass is accurately weighed and packaged in a standard crucible. According to the matched analysis procedure, differential scanning calorimetry is performed on each packaged sample to obtain the original heat flow curve. For each original heat flow curve, integral calculation is performed under a unified baseline to extract the gelatinization enthalpy value corresponding to each sample. The coordinates and risk level of the sample to which the gelatinization enthalpy value belongs are then identified to form a thermodynamic dataset.

6. The system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage according to claim 1, characterized in that, The formation process of the structured spatial sample database is as follows: for each sample in the rice sample set, a unique identifier including spatial coordinates is created; using the unique identifier as a clue, the risk level of matching coordinates is found in the quality deterioration risk map, and the gelatinization enthalpy value of matching identifiers is extracted from the thermodynamic dataset; the spatial coordinates, risk level, gelatinization enthalpy value and unique identifier corresponding to each sample are written into the database table as a complete record. After traversing all samples and completing the record writing, the consistency of all records is verified to generate the final structured spatial sample database.

7. The system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage according to claim 1, characterized in that, The process employs a covariate kriging algorithm, specifically including: calculating the spatial covariant relationship between the enthalpy of gelatinization and the risk value within the entire sample space based on a structured spatial sample database; constructing a kriging interpolation rule with spatial coordinates as the independent variable, the enthalpy of gelatinization as the dependent variable, and the risk value as the spatial drift auxiliary variable based on the spatial covariant relationship; calculating the positions of all unsampled points within the storage space according to the kriging interpolation rule, generating a predicted enthalpy of gelatinization and the corresponding kriging prediction variance for each point; integrating the predicted values ​​of all points into an aging degree prediction surface data, and integrating the corresponding prediction variances into prediction uncertainty assessment data.

8. The system for monitoring and analyzing the processing characteristics of early indica rice during storage for rice flour production, as described in claim 7, is characterized in that... The calculation of the spatial covariant relationship between the enthalpy of gelatinization and the risk value in the entire sample space specifically includes: determining multiple different spatial lag distance groups, and dividing the spatial coordinates of all samples in the structured spatial sample database into corresponding groups according to their mutual distances; for each spatial lag distance group, calculating the mean square of the difference between the enthalpy of gelatinization values ​​and the mean square of the difference between the risk values ​​of all sample pairs within the group, and obtaining the semivariance of the enthalpy of gelatinization and the semivariance of the risk value corresponding to the group; based on all spatial lag distance groups and the corresponding two sets of semivariance data, respectively fitting and generating a semivariance function of the enthalpy of gelatinization describing the spatial autocorrelation characteristics of the enthalpy of gelatinization, and a semivariance function of the risk value describing the spatial autocorrelation characteristics of the risk value.

9. The system for monitoring and analyzing the processing characteristics of early indica rice used for rice flour storage according to claim 1, characterized in that, The process of generating a quality decision report specifically includes: calculating a confidence interval for the predicted value of each point in the aging degree prediction surface data based on the Kriging prediction variance in the prediction uncertainty assessment data; identifying and delineating high-confidence local aging peak areas in the aging degree prediction surface data where the predicted value exceeds a preset quality threshold and the lower limit of the confidence interval also exceeds the corresponding threshold; statistically analyzing the distribution of predicted values ​​for all points in the aging degree prediction surface data, calculating the percentile-corresponding predicted values ​​as the overall warehouse aging degree statistical interval, and simultaneously calculating the total volume and average predicted confidence probability of the high-confidence local aging peak areas; and integrating the three-dimensional visualized starch aging spatial distribution map, the spatial coordinate range of the high-confidence local aging peak areas, the overall warehouse aging degree statistical interval, the total volume, and the average predicted confidence probability to generate a quality decision report.

Images