Method and system for analyzing changes in potato storage quality

Abstract

This invention discloses a method and system for analyzing changes in the storage quality of potatoes, relating to the field of food storage technology. The method includes: assigning unique identifiers to potato samples to be analyzed and constructing an inherent anchor point benchmark database; generating anchor point homogeneous time-series datasets with storage duration as the time-series axis for potato samples with uniform identifiers at preset fixed time intervals; executing a spatiotemporal anchoring homogeneous matching algorithm to extract the specific degradation time-series difference and degradation spatial diffusion gradient time-series sequence of degradation points; solving to generate a full-cycle spatiotemporal coupled degradation surface of the potato's true degradation process from local to overall; and outputting an analysis report on the spatiotemporal evolution law of the storage quality of the corresponding potato samples. This invention can accurately reconstruct the entire process of quality degradation, identify key inflection points of degradation, provide accurate data support for real-time control of storage processes, and effectively reduce post-harvest storage losses of potatoes.

Description

Technical Field

[0001] This invention relates to the field of food storage technology, and in particular to a method and system for analyzing changes in the storage quality of potatoes. Background Technology

[0002] Potatoes are my country's fourth largest food crop, with the world's largest annual output. During post-harvest storage, they are prone to quality deterioration problems such as sprouting, accumulation of reducing sugars, nutrient degradation, and disease growth, resulting in serious resource waste and economic losses. Accurate analysis of changes in storage quality is a key technical link to reduce post-harvest losses and ensure stable supply in the industrial chain.

[0003] Existing potato storage quality analysis technologies are mainly divided into two categories: traditional physicochemical benchmark testing and modern non-destructive testing. Traditional physicochemical testing provides accurate and reliable results and is the industry-recognized gold standard for quality evaluation. However, it is a destructive test that requires irreversible pretreatment of the sample and cannot continuously track and test the same sample throughout the entire storage cycle. While existing non-destructive testing technologies have solved the pain points of destructive testing, they have a single core technical deficiency: they can only perform static quality testing at a single storage time point for potatoes. They lack the ability to dynamically and continuously track the quality changes of the same sample throughout the entire storage cycle and to systematically analyze the patterns of quality changes. They cannot accurately reconstruct the dynamic quality deterioration process of a single sample during storage, and it is difficult to identify key inflection points of quality deterioration. Therefore, they cannot provide timely and accurate data support for real-time control of storage processes.

[0004] Therefore, developing a method that enables continuous dynamic tracking of the same sample throughout the entire potato storage cycle and precise analysis of quality change patterns is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] This invention provides a method for analyzing changes in the storage quality of potatoes, including: Step 1: Assign a unique identifier to each potato sample to be analyzed and construct its inherent anchor benchmark database; Step 2: At a preset fixed time interval, perform continuous, non-destructive hyperspectral imaging detection on potato samples with unified identification, and simultaneously collect real-time storage environment parameter sets at corresponding time nodes to generate anchor point homogeneous time series datasets with storage duration as the time axis. Step 3: Based on the inherent anchor point benchmark database and anchor point homogeneous time series dataset of the analysis samples, execute the spatiotemporal anchor homogeneous matching algorithm to extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of the degradation point. Step 4: Based on the specific degradation time difference of potato degradation sites and the degradation spatial diffusion gradient time sequence, solve to generate the full-cycle spatiotemporal coupled degradation surface of the potato from local to global real degradation process. Step 5: Based on the generated full-cycle spatiotemporal coupled deterioration surface, output an analysis report on the spatiotemporal evolution of storage quality of the corresponding potato sample.

[0006] The method for analyzing changes in potato storage quality as described above, which involves constructing an inherent anchor point benchmark database for potato analysis samples, specifically comprises the following sub-steps: Initial hyperspectral three-dimensional data of the samples were acquired and analyzed using a hyperspectral imaging system; The four types of inherent anatomical acupoints were located on the analysis sample, and the contour range of each site was stored to generate a permanent set of spatial anchor points. Extract the initial hyperspectral 3D data within the contour range of each site, process it into spectral features, and then bind its site number and analysis sample ID; Tissue samples were taken from the four types of inherent anatomical acupoint regions on parallel samples for testing, and the obtained test results were bound to the corresponding acupoint numbers and analysis sample IDs. Collect and analyze the initial parameter set of the sample storage environment, and record the collection timestamp as the starting point for timing the sample storage duration. Using the sample ID as the core index, the initial hyperspectral three-dimensional data, permanent spatial anchor set, site spectral features, site tissue detection results, and initial parameter set of storage environment are structurally correlated to construct a benchmark database specific to the sample.

[0007] The method for analyzing changes in potato storage quality described above, which generates a homogeneous time-series dataset anchored by storage duration as the time-series axis, specifically consists of the following sub-steps: A full-cycle time-series acquisition scheme is pre-designed for the analysis samples; The analysis samples were subjected to standardized acquisition of hyperspectral three-dimensional data according to the full-cycle time-series acquisition scheme, and real-time storage environment parameters were recorded simultaneously. Using the sample ID as the core index and the actual storage duration as the time series axis, the hyperspectral three-dimensional data and storage environment parameters collected in this study are structurally correlated to generate an anchor point homogeneous time series dataset.

[0008] The potato storage quality change analysis method described above, wherein the spatiotemporal anchoring homology matching algorithm is specifically divided into the following sub-steps: Using the set of permanent spatial anchors in the intrinsic anchor reference database as the unique spatial reference, spatial anchors are locked for hyperspectral images in the anchor-originating time-series dataset. For each locked spatial anchor point, extract the temporal feature sequence of its corresponding region; Using the initial reference hyperspectral feature set corresponding to each spatial anchor point as the unique temporal anchor point, baseline correction is performed on the temporal feature sequence of each anchor point to obtain a standardized anchor point quality temporal dataset; Based on a standardized anchor point quality time series dataset, we calculate and extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of degradation points.

[0009] The method for analyzing changes in potato storage quality, as described above, involves generating a full-cycle spatiotemporally coupled degradation surface representing the actual degradation process from local to overall degradation, based on the specific degradation time difference and the spatial diffusion gradient time sequence of the degradation sites. This process is further divided into the following sub-steps: Based on the specific degradation time difference set of degradation sites, the starting point of full-cycle degradation analysis is determined; For each spatial anchor point, a dedicated deterioration dynamics sub-model is constructed, and the deterioration time series curve of each anchor point is generated by solving the model. The degraded coupling effect between spatial anchor points is quantified, and a spatiotemporal coupling matrix is ​​constructed. Using the degradation time series curve of the starting point as the driving term, and based on the exclusive degradation dynamics sub-model of all spatial anchor points, a global degradation dynamics model with spatial diffusion-temporal evolution dual coupling is constructed by combining the spatiotemporal coupling matrix and the degradation spatial diffusion gradient time series sequence. The degradation degree distribution of the analysis sample in the continuous spatiotemporal domain is obtained by solving the model. Based on the analysis of the degradation distribution in the continuous spatiotemporal domain of the sample, a full-cycle spatiotemporal coupled degradation surface is generated.

[0010] This invention also provides a potato storage quality change analysis system, characterized in that it includes: The benchmark database construction module is used to assign a unique identifier to each potato sample to be analyzed and to build its own anchor benchmark database. The homogeneous time series dataset acquisition module is used to perform continuous, non-destructive hyperspectral imaging detection on potato samples with unified identification at preset fixed time intervals, and simultaneously collect the real-time storage environment parameter set at the corresponding time node to generate a homogeneous time series dataset with storage duration as the time axis. The homogeneous matching module is used to perform a spatiotemporal anchoring homogeneous matching algorithm based on the inherent anchor benchmark database and the anchor homogeneous time series dataset of the analysis samples, and extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of the degradation point. The degradation surface solution module is used to solve and generate the full-cycle spatiotemporal coupled degradation surface of potatoes from local to global real degradation process based on the specific degradation time difference and degradation spatial diffusion gradient time sequence of potato degradation points. The analysis report output module is used to output an analysis report on the spatiotemporal evolution of storage quality of the corresponding potato sample based on the generated full-cycle spatiotemporal coupled deterioration surface.

[0011] The beneficial effects achieved by this invention are as follows: It realizes continuous and non-destructive dynamic tracking of the same sample throughout the entire storage cycle of potatoes; through spatiotemporal anchoring homogeneous matching algorithm and spatiotemporal coupling deterioration dynamic modeling, it accurately restores the entire process of quality deterioration, identifies key deterioration inflection points, provides accurate data support for real-time control of storage processes, and effectively reduces post-harvest storage losses of potatoes. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0013] Figure 1 This is a flowchart of a method for analyzing changes in the storage quality of potatoes provided in Embodiment 1 of this application; Figure 2 This is a schematic diagram of a potato storage quality change analysis system provided in Embodiment 2 of this application. Detailed Implementation

[0014] 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, not all, of the embodiments of the present invention. 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.

[0015] Example 1

[0016] like Figure 1 As shown in Embodiment 1 of this application, a method for analyzing changes in the storage quality of potatoes is provided, comprising: Step S10: Assign a unique identifier to each potato sample to be analyzed and construct its inherent anchor benchmark database; Several potatoes of the same variety, harvested from the same batch, and at the same maturity were selected as analysis samples, and a parallel sample of the same size, batch, and quality was prepared for each analysis sample. Each analysis sample was assigned a unique QR code identification, engraved on the smooth, non-eye area of ​​the skin. Subsequently, a unique digital ID was assigned to each QR code identification, and the basic information of the corresponding sample, including variety, origin, and harvest time, was archived simultaneously, establishing a one-to-one binding relationship between the digital ID, QR code, and sample basic information to ensure traceability and no confusion of sample identity throughout the entire testing cycle. Then, the following testing procedure was performed on each analysis sample to construct its inherent anchor benchmark database: Step S11: Acquire and analyze the initial hyperspectral three-dimensional data of the sample using a hyperspectral imaging system; Turn on the hyperspectral imaging system and preheat the dual-branch halogen linear light source and spectrometer for 30 minutes in advance to eliminate light intensity fluctuations and dark current noise of the instrument. The system parameters are fixed as follows: wavelength coverage range of 400~1000nm visible-near infrared band, spectral resolution of 2.8nm, spatial resolution of 0.12mm / pixel, and scanning step size of 0.05mm. The analysis sample is fixed on a high-precision electronically controlled displacement stage, the fixed lens is perpendicular to the equatorial plane of the sample, the object distance is constant at 300mm, the dual linear light sources are symmetrically distributed on both sides of the lens, and the incident angle of the light sources is fixed at 45° to ensure that the analysis sample is uniformly illuminated without shadows. First, standard whiteboard calibration and full black background calibration are completed to eliminate system noise and ambient light interference. Then, the equatorial plane of the analysis sample is continuously scanned in 360° in 4 orthogonal planes, and the data is stitched together to generate an initial hyperspectral 3D dataset covering the entire surface and internal transmission information of the analysis sample. The dataset contains two-dimensional image information in the spatial dimension and one-dimensional full-band spectral information in the spectral dimension. Record the geometric pose parameters, light source parameters, ambient temperature and humidity parameters, and acquisition timestamp of this acquisition as the sole standardized acquisition conditions for subsequent full-cycle time-series data acquisition, ensuring that all data from the same source throughout the entire cycle.

[0017] Step S12: Locate the four types of inherent anatomical acupoints on the analysis sample, and store the contour range of each site to generate a permanent spatial anchor point set; Based on the initial hyperspectral 3D data collected, combined with the anatomical features of potatoes, and through hyperspectral feature matching and image contour recognition algorithms, the system accurately identified and located four types of inherent anatomical acupoints on the analyzed samples at the subpixel level. Bud eye sites: By analyzing the contour depression features and spectral reflectance differences of the sample epidermis, all natural bud eye areas of the epidermis are identified. Each complete bud eye is marked as an independent analysis unit, and the contour range and three-dimensional spatial coordinates of each bud eye are recorded. Vascular bundle ring sites: Using hyperspectral transmission scanning data, identify and analyze the characteristic regions of vascular bundle rings in the cross-section of the sample, and uniformly mark no less than 8 equidistant vascular bundle ring sites along the circumferential direction, recording the contour range and three-dimensional spatial coordinates of each site; No-damage sites on the epidermis: In the smooth area of ​​the epidermis of the analysis sample without buds, damage, or lesions, no less than 6 no-damage sites on the epidermis are uniformly marked, with a distance of no less than 10 mm between adjacent sites, and the contour range and three-dimensional spatial coordinates of each site are recorded. Illness location: Identify and analyze the core region of the medulla at the center of the cross-section of the sample, mark one medulla center location, and record the contour range and three-dimensional spatial coordinates of the location; All identified and located anatomical sites are assigned a unique numerical number, following the rule of [sample ID - site type code - site sequence number]. A permanent set of spatial anchor points is generated, indexed by the site number and containing the site's three-dimensional spatial coordinates, site type, site outline range, and the analysis sample ID to which the site belongs. This ensures that throughout the entire storage period, spatial anchor points with the same number always correspond to the same anatomical location on the analysis sample.

[0018] Step S13: Extract the initial hyperspectral 3D data within the contour range of each site, process it into spectral features, and then bind its site number and analysis sample ID; Based on the generated set of permanent spatial anchor points, for each numbered spatial anchor point (i.e., anatomical acupoint), the hyperspectral raw data corresponding to all pixels within the anchor point contour range is accurately extracted; To address the differences in optical properties among different types of anatomical acupoints, a feature extraction operator specific to each anchor point is invoked to extract the initial hyperspectral feature set corresponding to each anchor point, including the average reflectance across the entire band, quality-related wavelength reflectance, spectral first derivative features, spectral absorption depth features, and scattering feature values. Background signals outside the anchor point region, cross-interference signals from adjacent tissues, and abnormal noise spectra are removed to generate an independent, interference-free initial hyperspectral feature dataset for each spatial anchor point, which is then bound to the corresponding site number and analysis sample ID.

[0019] Step S14: Sample and test the tissues in the four types of inherent anatomical acupoint areas on the parallel samples, and bind the obtained test results with the corresponding acupoint number and analysis sample ID; Using parallel samples from the same batch as the analysis samples, and following the site location rules in step S13, the same type and number of anatomical acupoints were marked on the parallel samples, and the tissue corresponding to each marked site was precisely sampled. According to the current national standard methods, the core quality physicochemical benchmark values ​​of each sampling site were determined, including: the pulp firmness was determined by texture analyzer, the reducing sugar content was determined by Fehling's reagent titration method, the starch content was determined by acid hydrolysis method, and the vitamin C content and solanine content were determined by high performance liquid chromatography. The measured physicochemical benchmark values ​​are bound one by one with the corresponding site type, site number, and analysis sample ID to generate an anchor point tissue detection result set.

[0020] Step S15: Collect and analyze the initial parameter set of the sample storage environment, and record the collection timestamp as the starting point for timing the sample storage duration; The analysis samples are transferred to a fixed placement area in the target storage environment, and multi-parameter environmental acquisition nodes are simultaneously deployed in the sample placement area. The node sampling frequency is matched with the time interval of subsequent time series data acquisition. The initial parameter set of the sample storage environment is collected and analyzed, including real-time ambient temperature, relative humidity, carbon dioxide concentration, and oxygen concentration. The initial collection timestamp of all parameters is recorded, and this timestamp is marked as the starting point for the storage duration of the sample.

[0021] Step S16: Using the sample ID as the core index, the initial hyperspectral three-dimensional data, permanent spatial anchor set, site spectral features, site tissue detection results, and initial storage environment parameter set are structurally correlated to construct a benchmark database specific to the sample. An immutable encrypted storage format was used to construct a benchmark database dedicated to the analysis sample, which serves as the sole benchmark for subsequent full-cycle time-series data source matching, benchmark correction, and quality degradation analysis.

[0022] Step S20: At a preset fixed time interval, perform continuous, non-destructive hyperspectral imaging detection on potato samples with unified identification, and simultaneously collect the real-time storage environment parameter set at the corresponding time node to generate a homogeneous time-series dataset with storage duration as the time axis. The rules, terminology, and data format for collecting anchor point time series data are strictly consistent with the benchmark database in step S10 to ensure that the full-cycle time series data and the benchmark data are comparable. This is specifically divided into the following sub-steps: Step S21: Pre-set a full-cycle time-series acquisition scheme for the analysis samples; Based on the varietal characteristics, target storage period, and storage mode (cellar storage / ventilated storage / cold storage / controlled atmosphere storage) of the analyzed samples, a fixed collection time interval is preset: 3 to 7 days under conventional storage scenarios. For varieties prone to germination or high-temperature and high-humidity storage environments, a 7-day interval can be set during the dormancy period, a 3-day interval during the germination induction period, and a 1-day interval during the germination acceleration period, forming a phased differentiated collection rule. An automated collection trigger rule is established, with the storage duration time starting point calibrated in step S10 as the sole benchmark. When the actual storage duration reaches an integer multiple of the preset time interval, the collection process is automatically triggered.

[0023] Step S22: Perform standardized acquisition of hyperspectral three-dimensional data of the analysis sample according to the full-cycle time-series acquisition scheme, and simultaneously record real-time storage environment parameters; After the data acquisition process is triggered, the analysis sample with the corresponding identification is retrieved from the fixed placement area in the storage environment. The QR code identification on the sample surface is scanned using an industrial barcode scanner to complete the sample ID verification. Then, the sample is fixed on a high-precision electrically controlled displacement stage. Using the reference acquisition geometric pose parameters recorded in step S10 as the sole standard, the three-dimensional pose of the analysis sample is precisely reset. A scanning process identical to step S10 is then executed, performing a continuous linear array scan of the analysis sample's equatorial plane in four orthogonal planes at a 360° angle. The scanning start position and scanning range for each orthogonal plane are specified. The scanning step size is perfectly matched with the baseline acquisition, with no scanning areas omitted or repeated. The raw scanning data from the four orthogonal planes are stitched together without distortion to generate a time-series hyperspectral 3D dataset covering the entire epidermis and internal transmission information of the analyzed sample at this time point. The dimensions, format, and storage rules of the dataset are completely consistent with the initial hyperspectral 3D dataset acquired in step S10. The precise timestamp of this acquisition, the corresponding actual storage time, and the temperature and humidity parameters of the acquisition environment are recorded synchronously and bound one-to-one with the hyperspectral 3D dataset at this time point to ensure data traceability.

[0024] Within the same time window of hyperspectral imaging data acquisition, the real-time parameter set of the storage environment where the analysis sample is located is synchronously acquired through the multi-parameter environmental acquisition nodes deployed in step S10. The parameter types are completely consistent with the initial storage environment parameter set. The sampling frequency of environmental parameters is not less than 1 time / minute. The average value of the parameters within the entire time window of hyperspectral acquisition is taken as the effective environmental parameter value of this time series node to eliminate parameter deviations caused by instantaneous environmental fluctuations. The real-time storage environment parameter set of this time series node is precisely bound to the hyperspectral three-dimensional dataset of the same node, the unique ID of the analysis sample, the acquisition timestamp, and the actual storage duration to ensure the temporal homology between environmental parameters and quality detection data, providing accurate independent variable data for the subsequent deterioration kinetic model.

[0025] Step S23: Using the analysis sample ID as the core index and the actual storage duration as the time series axis, the hyperspectral three-dimensional data and storage environment parameters collected this time are structured and correlated to generate an anchor point homogeneous time series dataset. Using the unique digital ID of the analysis sample as the core index and the actual storage duration as the time series axis, the hyperspectral 3D dataset collected this time, the real-time storage environment parameter set collected synchronously, and the collection timestamp are structured and correlated. The structured time series node data is then correlated and matched with the analysis sample-specific benchmark database constructed in step S10 to ensure that the data of each time series node can be accurately mapped to the permanent spatial anchor set in the benchmark database, generating the anchor homologous time series data unit of that time series node. The association and matching rules between the time-series node data and the dedicated benchmark database include primary sample identity matching and secondary site precise matching. Primary sample identity matching refers to using the unique digital ID of the analyzed sample as the unique matching key to bind the time-series node data collected this time to the benchmark database with the same ID that was built before storage. Secondary site precise matching refers to, based on sample ID matching, using a spatial coordinate mapping algorithm to precisely align the pixel coordinates in the newly collected hyperspectral 3D data with the 3D spatial coordinates of permanent anchor points in the benchmark database at the sub-pixel level, so that a certain pixel area in the newly collected data corresponds 1:1 precisely to a certain numbered anchor point in the benchmark data.

[0026] According to the time sequence of storage duration, the anchor-source time series data units of all time series nodes in the entire cycle are archived in an orderly manner, and finally a complete anchor-source time series dataset with storage duration as the time sequence axis is formed.

[0027] Step S30: Based on the inherent anchor point benchmark database and anchor point homogeneous time series dataset of the analysis sample, execute the spatiotemporal anchor homogeneous matching algorithm to extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of its degradation point. The spatiotemporal anchoring homology matching algorithm specifically includes the following sub-steps: Step S31: Using the permanent spatial anchor set in the inherent anchor reference database as the unique spatial reference, perform spatial anchor locking on the hyperspectral images in the anchor-originating time series dataset; From the inherent anchor point benchmark database specific to the analysis sample, the permanent spatial anchor point set of the sample is retrieved, and the initial three-dimensional spatial coordinates, contour range, and site type code of each spatial anchor point are obtained to generate a unique benchmark template for the entire life cycle anchoring of the sample; at the same time, the anchor point homogeneous time series data units of each time series node in the anchor point homogeneous time series data are retrieved in sequence. For each spatial anchor point in the permanent spatial anchor point set, a gradient-direction-based sub-pixel contour matching algorithm is used to accurately map the contour range of the anchor point in the reference template to the hyperspectral image coordinate system of each time node, locking the precise contour range and sub-pixel-level three-dimensional spatial coordinates of the anchor point at each time node; for tiny feature sites such as bud eyes and vascular bundle rings, a local feature enhancement algorithm is used to amplify the contour features, ensuring that the site contour locking accuracy is not less than 0.05mm; For all locked valid anchor points, their contour range, three-dimensional spatial coordinates, and site number are structured and archived, and updated to the anchor point homogeneous time series data unit of the corresponding time series node. This serves as the sole spatial range basis for subsequent anchor point feature extraction, ensuring that spatial anchor points with the same number always correspond to the same anatomical location on the analysis sample throughout the entire cycle.

[0028] Step S32: For each locked spatial anchor point, extract the temporal feature sequence of its corresponding region; For each locked and valid spatial anchor point, based on its archived contour range and spatial coordinates, the raw hyperspectral data corresponding to all valid pixels within the contour range is accurately extracted. Background pixels, edge transition pixels, and cross-interference pixels from adjacent tissues outside the contour range are removed, retaining only the pure hyperspectral data of the core region of the anchor point to avoid signal interference from non-target regions. The anchor point-specific feature extraction operators pre-constructed in step S13, which correspond one-to-one with the site type, are called to extract features from the pure hyperspectral data of each anchor point. The operator matching rules and extraction dimensions are as follows: Bud eye sites: Prioritize the extraction of characteristic wavelength spectral features related to germination and hormone metabolism, including characteristic wavelength reflectance, spectral first derivative features, and scattering characteristic values; Vascular bundle ring sites: Prioritize the extraction of characteristic wavelength spectral features related to nutrient transport, reducing sugar accumulation, and starch degradation, including characteristic wavelength reflectance, spectral absorption depth features, and second derivative features; Undamaged epidermal sites: Prioritize the extraction of characteristic wavelength spectral features related to epidermal dehydration, disease infection, and epidermal browning, including the average reflectance across the entire wavelength band, the first derivative of the spectrum, and scattering characteristic values. Bone marrow location: Prioritize extraction of characteristic wavelength spectral features related to overall aging, overall degradation of nutrients, and enrichment of solanine, including characteristic wavelength reflectance, spectral absorption depth features, and average reflectance across the entire wavelength range; The extracted anchor point original hyperspectral features are smoothed using a 5-point second-order Savitzky-Golay filtering algorithm to eliminate spectral random noise. Outliers in the feature sequence are removed using the 3σ criterion, and the removed outliers are supplemented by interpolation using the feature mean of adjacent time nodes to ensure the continuity and smoothness of the feature sequence. Following the temporal order of storage duration, the denoised hyperspectral features of spatial anchor points with the same number across all temporal nodes in the entire period are arranged in an ordered manner to generate an independent anchor point temporal feature sequence that corresponds one-to-one with the storage duration. The mathematical expression for any temporal feature in this sequence is defined as follows: Where k is the spatial anchor index and t is the actual storage duration of the corresponding time-series node. Let be the total number of hyperspectral feature dimensions of the k-th spatial anchor point. For the k-th spatial anchor point at storage time t, the first... Spectral features.

[0029] Finally, the time series feature sequences of all anchor points are structurally bound with the corresponding anchor point number and analysis sample ID, and updated to the anchor point source time series data unit.

[0030] Step S33: Using the initial reference hyperspectral feature set corresponding to each spatial anchor point as the unique temporal anchor point, perform baseline correction on the temporal feature sequence of each anchor point to obtain a standardized anchor point quality temporal dataset; For the k-th spatial anchor point, the spectral features of its associated sites are retrieved from the intrinsic anchor point reference database and used as the unique time-series anchor point for full-cycle correction of that anchor point, denoted as: ,in For the k-th spatial anchor point at the initial time (t=0), the k-th time... Spectral features; Based on the initial hyperspectral feature set of the anchor points obtained in step S13 and the tissue detection result set of the anchor points obtained in step S14, the existing technical method of quantitative analysis of Pearson correlation coefficient + significance test is adopted to predefine two exclusive feature dimension sets: ① quality-related feature dimension set ① A set of spectral feature dimensions strongly correlated with potato flesh firmness, reducing sugar content, starch content, vitamin C content, and solanine content; ② A set of non-quality reference feature dimensions. The set of spectral feature dimensions, which are not affected by changes in the intrinsic quality of potatoes but are only sensitive to environmental fluctuations and equipment drift, serves as the calculation benchmark for baseline drift. At storage duration t, based on a unique time-series anchor point Quality-related feature dimension set Non-quality reference feature dimension set Calculate the systematic baseline offset coefficient for each anchor point. The calculation formula is: ,in Indexed by spectral feature dimensions. , For the k-th spatial anchor point at storage time t, the first... Spectral features, For the k-th spatial anchor point at the initial time (t=0), the k-th time... Spectral features; Based on drift coefficient With a unique timing anchor Calculate the anchor point at storage time t. Dynamic baseline values ​​of each spectral feature The calculation formula is: ; Based on dynamic baseline value For each time-series feature in the original time-series feature sequence The correction is performed using the following formula: ,in For the k-th spatial anchor point at storage time t, the first... spectral features The corrected value These are the weight coefficients for feature dimensions, specifically for quality-related feature dimensions. , For non-quality reference feature dimensions , For the remaining feature dimensions ; After performing the above corrections on all time-series features of all anchor points, a standardized anchor point quality time-series dataset is constructed using the anchor point number as the index and the storage duration as the time-series axis. This dataset is then structurally bound to the analysis sample ID and anchor point number.

[0031] Step S34: Based on the standardized anchor point quality time series dataset, calculate and extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of the degradation point; First, calculate the overall quality deterioration of the kth spatial anchor point at storage time t. The calculation formula is: ,in This represents the total number of quality-related feature dimensions. In the anchor quality time series dataset, the k-th spatial anchor at storage duration t is the first spatial anchor. Spectral features, , A set of quality-related feature dimensions; The value range is [0,1]. This means that the k-th spatial anchor point is completely consistent with the initial baseline quality, without any deterioration. The closer the value is to 1, the more severe the quality deterioration of the anchor point. Subsequently, the degradation threshold is determined based on the preset value. Determine the deterioration start time of the anchor point, i.e., when Two consecutive time-series nodes continuously exceed the threshold When the anchor point is determined to have entered a deterioration state, the corresponding storage duration is recorded as the deterioration start time of the anchor point. ; Traversing the degradation start time of all anchor points Extract the minimum value, denoted as . , The corresponding anchor index is denoted as For any spatial anchor point, using the formula: Calculate its specific degradation time difference ,in Anchor point type-specific weighting coefficients are predefined based on the deterioration sensitivity of different anatomical acupoints in potatoes: eye sites. Vascular bundle ring site No lesions on the epidermis Bone marrow location ; Put all anchor points , It constitutes a set of site-specific deterioration time differences, which fully characterizes the order and time difference of deterioration at different anatomical sites, and completes the structured binding with the analysis sample ID and corresponding anchor point number; definition The three-dimensional spatial coordinates are The three-dimensional spatial coordinates of any adjacent anchor points are Calculate the Euclidean spatial distance between the two. The unit is mm, representing the spatial distance of degradation diffusion, m is The adjacent anchor index; for any storage duration t, using the formula: Calculate the deterioration path →m direction diffusion gradient component ,in For the first The overall quality deterioration at each anchor point at storage time t. The overall quality deterioration degree of the m-th anchor point at storage time t; further calculate and analyze the time series value of the spatial diffusion gradient of deterioration of the sample at storage time t. The calculation formula is: Where M is the first The number of adjacent anchor points around each anchor point The diffusion direction weighting coefficient is predefined based on the potato's anatomical structure and the law of deterioration diffusion: anchor points along the vascular bundle ring direction. Anchor points along the radial skin direction Anchor points along the direction of the medulla oblongata Finally, according to the chronological order of storage duration, all time-series nodes of the entire cycle are... Arrange them in an orderly manner to generate a time series sequence of degraded spatial diffusion gradients.

[0032] The calculated site-specific degradation time series difference set and degradation spatial diffusion gradient time series are structurally bound with the unique ID of the analysis sample, the corresponding anchor point number, and the storage duration.

[0033] Step S40: Based on the specific degradation time difference of potato degradation sites and the degradation spatial diffusion gradient time sequence, solve to generate the full-cycle spatiotemporal coupled degradation surface of the real degradation process of potatoes from local to global. A site-specific spatiotemporal coupled degradation dynamics model is constructed to solve for and generate the full-cycle spatiotemporal coupled degradation surface of the potato from local to global degradation process. The specific steps include the following: Step S41: Based on the specific degradation time series difference set of degradation sites, determine the starting site for full-cycle degradation analysis; Traverse the degradation start time of all spatial anchors in the specific degradation time difference set. The anchor point corresponding to the minimum value is taken as the starting point for the full-cycle degradation analysis. The time of onset of its deterioration is denoted as The three-dimensional spatial coordinates are as follows ,Will The zero point is calibrated as the time-series baseline for the full-cycle deterioration evolution. The reference origin is defined as the diffusion point of the deteriorated space; with Using the standardized anchor quality time series dataset as the driving benchmark, all secondary sites (at) are collected. The standardized anchor quality time series dataset (other anchor points are called secondary sites), the corresponding storage environment parameter set, and the deterioration spatial diffusion gradient time series sequence form the basic dataset for deterioration dynamics modeling.

[0034] Step S42: For each spatial anchor point, construct its own deterioration dynamics sub-model and solve to generate the continuously differentiable site deterioration time series curve for each anchor point; For storage duration t, construct the environmental cumulative effect function for the k-th spatial anchor point. To address the shortcomings of existing technologies that only use instantaneous environmental parameters for modeling and cannot reconstruct long-term cumulative environmental damage, the expression is: ,in The sampling interval for environmental parameters. This represents the total number of sampling points from the start of storage (corresponding to i=0, t=0) to the target time t. , These are the measured values ​​of the j-th environmental parameter (temperature / relative humidity / carbon dioxide concentration / oxygen concentration) at the i-th and (i-1)-th sampling times, respectively. This represents the optimal value (system preset value) for the j-th environmental parameter. The sensitivity coefficient of the k-th spatial anchor point to the j-th environmental parameter is based on the predefined site type (the sensitivity coefficient of the bud site to temperature). Sensitivity coefficient of vascular bundle ring sites to gas concentration Epidermal site sensitivity to humidity The overall sensitivity coefficient of the bone marrow site to temperature and oxygen ); Based on the first-order kinetic mechanism of degradation reaction, and combining the cumulative effect of the environment and the site degradation characteristics, a differential expression for a specific degradation kinetic sub-model is constructed: ,in Let be the degradation rate of the k-th spatial anchor point (the first derivative of the overall degradation with respect to time). Let be the overall quality deterioration degree of the k-th spatial anchor point at storage time t. The intrinsic degradation rate constant of the k-th anchor point (based on site type pre-calibration). This is the amplification factor for the cumulative environmental effect. The site's own aging coefficient characterizes the rate of natural aging and deterioration without environmental stress. The overall quality degradation rate output in step S34 For the training set, the optimal parameters of the sub-model are obtained by fitting the data using the nonlinear least squares method, resulting in a continuous analytical solution for the overall quality deterioration of the k-th spatial anchor point with respect to the storage duration t. That is, the continuously differentiable site degradation time-series curve of the anchor point, realizing the characterization of continuous degradation throughout the entire cycle.

[0035] Step S43: Quantify the degraded coupling effect between spatial anchor points and construct the spatiotemporal coupling matrix; The degradation coupling effect between spatial anchors includes: driving coupling effect and degradation initiation site. The deterioration-promoting effect at each secondary site is caused by the transport and diffusion of deteriorating metabolites along anatomical structures; the synergistic coupling effect, the synergistic deterioration-promoting effect between secondary sites, is caused by the cross-diffusion of deteriorating metabolites from adjacent sites and the synergistic effect of tissue aging.

[0036] The quantification formula for the driving coupling effect coefficient is expressed as: ,in The starting site The driving coupling effect coefficient for the kth secondary site ( ), Anatomical connectivity weights are predefined based on site type and potato material transport patterns (vascular bundle loop sites). , bud eye site Bone marrow location No lesions on the epidermis ), Spatial attenuation coefficient (preset to 0.08mm) -1 ), The starting site Euclidean spatial distance to the k-th secondary site, For the time difference of specific deterioration at the k-th secondary site, The starting site The time of onset of degradation; The quantitative formula for the synergistic coupling effect coefficient is expressed as: ,in For the kth, the Cooperative coupling effect coefficient between secondary sites ( , , ), The synergistic effect attenuation coefficient (preset to 0.6). For the kth, the Type matching weights between secondary sites (sites of the same type) Cross-type sites ), For the kth, the Euclidean spatial distance between secondary sites; Based on the coupling effect coefficients among all sites, construct The spatiotemporal coupling matrix C (where n is the total number of spatial anchor points) is defined by the following matrix elements: .

[0037] Step S44: Using the degradation time series curve of the starting site as the driving term, and based on the exclusive degradation dynamics sub-model of all spatial anchor points, combined with the spatiotemporal coupling matrix and the degradation spatial diffusion gradient time series sequence, construct a global degradation dynamics model with dual coupling of spatial diffusion and temporal evolution, and solve to obtain the degradation degree distribution of the analysis sample in the continuous spatiotemporal domain. The constructed global deterioration dynamics model is expressed as follows: ,in To analyze the overall quality deterioration degree of any spatial location P on the sample at storage time t, it is a continuous function of two variables, P and t, and serves as the objective of the model solution. For any location, the rate of degradation; , The inherent degradation rate constant and environmental effect amplification factor of location P are obtained by inverse distance weighted interpolation of the parameters of neighboring anchor points; The environmental cumulative effect value of location P is derived from its neighboring spatial anchors. The inverse distance weighted interpolation method is used to calculate D, which is the degradation diffusion coefficient, obtained from the degradation spatial diffusion gradient time series sequence in step S34. Fitting solution; The Laplace operator for the degree of deterioration characterizes the spatial diffusion effect of deterioration; The coupling effect coefficient between position P and the kth spatial anchor point is obtained by spatial interpolation of the spatiotemporal coupling matrix C; Let be the degradation rate of the k-th spatial anchor point, and n be the total number of spatial anchor points.

[0038] Set the boundary conditions for the model: Initial condition: at t=0 Boundary conditions: The degradation gradient of the epidermal boundary of the analyzed sample satisfies... ,in To analyze the sample epidermal boundary, For the boundary normal vector, This is the epidermal boundary mass transfer coefficient (pre-calibrated value). The average deterioration degree of the epidermal sites; anchor point constraint condition: at all anchor point locations , , The spatial position of the k-th spatial anchor point Overall quality deterioration at storage time t Let be the continuous analytical solution of the overall quality deterioration degree of the k-th spatial anchor point with respect to the storage time t.

[0039] The global spatiotemporal coupled model was discretized and solved using the finite volume method, with a spatial grid step size of 0.1 mm and a time step size of 0.5 days, to obtain the storage period. Within the analysis, the continuous degradation distribution at any location P in the entire spatial domain of the sample is analyzed. T represents the total duration of the target storage period.

[0040] Step S45: Based on the degradation degree distribution in the continuous spatiotemporal domain of the analysis sample, generate a full-cycle spatiotemporal coupled degradation surface; The three-dimensional spatial coordinates of the analysis sample Normalized to a two-dimensional parameterized plane , , It fully preserves the anatomical structure and spatial relative positions of potatoes; using normalized two-dimensional coordinates Using a planar coordinate system with storage duration t as the time series axis and deterioration degree as the coordinate axis. Using the vertical axis, a three-dimensional spatiotemporally coupled degradation surface is generated, which intuitively represents the degree of quality degradation at any storage time and any location of the potato, and completely restores the dynamic evolution process from the local starting point to the overall degradation of the whole potato. The discretized data of the degradation surface, the corresponding anchor point number, storage time, anatomical acupoint type, and environmental parameters are structured and bound, and archived with the unique ID of the analysis sample.

[0041] Step S50: Based on the generated full-cycle spatiotemporal coupled deterioration surface, output the spatiotemporal evolution analysis report of the storage quality of the corresponding potato sample; Using the sample ID as the core index, the following data are collected: the inherent anchor point benchmark database generated in step S10, the anchor point homogeneous time series dataset generated in step S20, the standardized anchor point quality time series dataset generated in step S33, the site-specific degradation time series difference set and degradation spatial diffusion gradient time series sequence generated in step S34, the full-cycle spatiotemporal coupled degradation surface and corresponding degradation degree distribution Q(P,t) generated in step S40, and the full-cycle storage environment parameter set, forming the analysis source data pool; all data follow the terminology system and data format of the aforementioned steps to complete the one-to-one matching of time series and spatial dimensions; Extracting core degradation characteristic parameters from the source data pool specifically includes: ① Extracting the degradation start time of each anchor point from the results of step S34. Specific deterioration timing difference Full-cycle deterioration spatial diffusion gradient time series Deterioration judgment threshold ; ② Extract and analyze the full-temporal and spatial distribution of degradation degree of the sample from the results of step S40. Continuous degradation time-series curves at each anchor point, degradation initiation sites Time reference zero point ③ Extract the cumulative environmental effect values ​​of each anchor point from the results of step S42. Complete the structured binding of all the above parameters with the corresponding anchor point number, storage duration, and analysis sample ID; Based on the extracted deterioration distribution Degraded spatial diffusion gradient time series Combined with the deterioration judgment threshold A three-tiered degradation inflection point determination rule is set: ① Early warning inflection point: The degradation degree of the first anchor point reaches 0.2× for two consecutive time series nodes. ,and ① First sustained positive growth; ② Rapid degradation inflection point: The spatial coverage of degradation in the analyzed sample reached 20%, and the average degradation degree of the entire sample exceeded the threshold for two consecutive time periods. ③ Severe deterioration inflection point: The spatial coverage of deterioration in the analyzed samples reaches 60%, and the average deterioration degree of the whole sample is ≥0.8; Identify the storage time, spatial location and environmental parameters at the triggering time corresponding to each inflection point, and complete the inflection point classification and calibration. Execution evolution pattern analysis: ① Time series dimension analysis: based on the deterioration start time of each anchor point ① Continuous degradation time-series curves are used to divide the entire storage cycle of the sample into quality stability period, early warning period, rapid degradation period, and failure period, clarifying the duration range and average degradation rate of each stage; ② Spatial dimension analysis: based on specific degradation time-series differences Deterioration distribution ③ Identify the initiation site and main diffusion path of sample deterioration; ④ Environmental coupling analysis: based on the cumulative environmental effect value. The influence weights of temperature, relative humidity, carbon dioxide concentration, and oxygen concentration on the sample deterioration rate were quantified to screen core environmental factors. The inflection point calibration results and evolution law analysis conclusions are combined with the original detection data from the source data pool to generate a standardized analysis report containing core data tables, deterioration characteristic curves, deterioration surface visualization maps, graded risk warning results, and phased storage process optimization suggestions. The generated report is then bound to the unique ID of the analysis sample, the QR code identification, and the full-cycle detection data, and archived using an tamper-proof encrypted format for easy access.

[0042] Example 2 like Figure 2 As shown, Embodiment 2 of this application provides a potato storage quality change analysis system, including: a benchmark database construction module 21, a homogeneous time series dataset acquisition module 22, a homogeneous matching module 23, a deterioration surface solution module 24, and an analysis report output module 25; The benchmark database construction module 21 is used to assign a unique identifier to each potato sample to be analyzed and to construct its inherent anchor benchmark database. The homogeneous time series dataset acquisition module 22 is used to perform continuous, non-destructive hyperspectral imaging detection on potato samples with unified identity at preset fixed time intervals, synchronously collect the real-time storage environment parameter set at the corresponding time node, and generate a homogeneous time series dataset with storage duration as the time axis. The homogeneous matching module 23 is used to perform a spatiotemporal anchoring homogeneous matching algorithm based on the inherent anchor benchmark database and the anchor homogeneous time series dataset of the analysis sample, and extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of its degradation point. The degradation surface solution module 24 is used to solve and generate the full-cycle spatiotemporal coupled degradation surface of the potato from local to global real degradation process based on the specific degradation time difference of the potato degradation point and the degradation spatial diffusion gradient time sequence. The analysis report output module 25 is used to output an analysis report on the spatiotemporal evolution of storage quality of the corresponding potato sample based on the generated full-cycle spatiotemporal coupled deterioration surface.

[0043] Corresponding to the above embodiments, the present invention provides a computer storage medium, including: at least one memory and at least one processor; The memory is used to store one or more program instructions; A processor is used to run one or more program instructions to execute a method for analyzing changes in potato storage quality.

[0044] Corresponding to the above embodiments, this embodiment of the invention provides a computer-readable storage medium containing one or more program instructions, which are executed by a processor to provide a method for analyzing changes in potato storage quality.

[0045] The embodiments disclosed in this invention provide a computer-readable storage medium storing computer program instructions. When the computer program instructions are executed on a computer, the computer performs the aforementioned method for analyzing changes in potato storage quality.

[0046] In this embodiment of the invention, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0047] The various methods, steps, and logic diagrams disclosed in the embodiments of this invention can be implemented or executed. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The processor reads information from the storage medium and, in conjunction with its hardware, completes the steps of the above methods.

[0048] The storage medium can be memory, such as volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.

[0049] Among them, non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory.

[0050] Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus RAM (DRRAM).

[0051] The storage media described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0052] Those skilled in the art will recognize that, in one or more of the examples above, the functions described in this invention can be implemented using a combination of hardware and software. When applied as software, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of computer programs from one place to another. Storage media can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0053] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for analyzing changes in the storage quality of potatoes, characterized in that, include: Step 1: Assign a unique identifier to each potato sample to be analyzed and construct its inherent anchor benchmark database; Step 2: At a preset fixed time interval, perform continuous, non-destructive hyperspectral imaging detection on potato samples with unified identification, and simultaneously collect real-time storage environment parameter sets at corresponding time nodes to generate anchor point homogeneous time series datasets with storage duration as the time axis. Step 3: Based on the inherent anchor point benchmark database and anchor point homogeneous time series dataset of the analysis samples, execute the spatiotemporal anchor homogeneous matching algorithm to extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of the degradation point. Step 4: Based on the specific degradation time difference of potato degradation sites and the degradation spatial diffusion gradient time sequence, solve to generate the full-cycle spatiotemporal coupled degradation surface of the potato from local to global real degradation process. Step 5: Based on the generated full-cycle spatiotemporal coupled deterioration surface, output an analysis report on the spatiotemporal evolution of storage quality of the corresponding potato sample.

2. The method for analyzing changes in potato storage quality according to claim 1, characterized in that, The construction of an inherent anchor benchmark database for potato analysis samples involves the following sub-steps: Initial hyperspectral three-dimensional data of the samples were acquired and analyzed using a hyperspectral imaging system; The four types of inherent anatomical acupoints were located on the analysis sample, and the contour range of each site was stored to generate a permanent set of spatial anchor points. Extract the initial hyperspectral 3D data within the contour range of each site, process it into spectral features, and then bind its site number and analysis sample ID; Tissue samples were taken from the four types of inherent anatomical acupoint regions on parallel samples for testing, and the obtained test results were bound to the corresponding acupoint numbers and analysis sample IDs. Collect and analyze the initial parameter set of the sample storage environment, and record the collection timestamp as the starting point for timing the sample storage duration. Using the sample ID as the core index, the initial hyperspectral three-dimensional data, permanent spatial anchor set, site spectral features, site tissue detection results, and initial parameter set of storage environment are structurally correlated to construct a benchmark database specific to the sample.

3. The method for analyzing changes in potato storage quality according to claim 1, characterized in that, Generating a homogeneous time-series dataset with storage duration as the time-series axis as the anchor point involves the following sub-steps: A full-cycle time-series acquisition scheme is pre-designed for the analysis samples; The analysis samples were subjected to standardized acquisition of hyperspectral three-dimensional data according to the full-cycle time-series acquisition scheme, and real-time storage environment parameters were recorded simultaneously. Using the sample ID as the core index and the actual storage duration as the time series axis, the hyperspectral three-dimensional data and storage environment parameters collected in this study are structurally correlated to generate an anchor point homogeneous time series dataset.

4. The method for analyzing changes in potato storage quality according to claim 1, characterized in that, The spatiotemporal anchoring homology matching algorithm is specifically divided into the following sub-steps: Using the set of permanent spatial anchors in the intrinsic anchor reference database as the unique spatial reference, spatial anchors are locked for hyperspectral images in the anchor-originating time-series dataset. For each locked spatial anchor point, extract the temporal feature sequence of its corresponding region; Using the initial reference hyperspectral feature set corresponding to each spatial anchor point as the unique temporal anchor point, baseline correction is performed on the temporal feature sequence of each anchor point to obtain a standardized anchor point quality temporal dataset; Based on a standardized anchor point quality time series dataset, we calculate and extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of degradation points.

5. The method for analyzing changes in potato storage quality according to claim 1, characterized in that, Based on the specific degradation time difference and degradation spatial diffusion gradient time sequence of potato degradation sites, the full-cycle spatiotemporal coupled degradation surface of the potato from local to global degradation process is generated. This is specifically divided into the following sub-steps: Based on the specific degradation time difference set of degradation sites, the starting point of full-cycle degradation analysis is determined; For each spatial anchor point, a dedicated deterioration dynamics sub-model is constructed, and the deterioration time series curve of each anchor point is generated by solving the model. The degraded coupling effect between spatial anchor points is quantified, and a spatiotemporal coupling matrix is ​​constructed. Using the degradation time series curve of the starting point as the driving term, and based on the exclusive degradation dynamics sub-model of all spatial anchor points, a global degradation dynamics model with spatial diffusion-temporal evolution dual coupling is constructed by combining the spatiotemporal coupling matrix and the degradation spatial diffusion gradient time series sequence. The degradation degree distribution of the analysis sample in the continuous spatiotemporal domain is obtained by solving the model. Based on the analysis of the degradation distribution in the continuous spatiotemporal domain of the sample, a full-cycle spatiotemporal coupled degradation surface is generated.

6. The method for analyzing changes in potato storage quality according to claim 1, characterized in that, Based on the generated full-cycle spatiotemporal coupled deterioration surface, an analysis report on the spatiotemporal evolution of storage quality of the corresponding potato sample is output, which is specifically divided into the following sub-steps: Construct an analysis source data pool and extract core degradation characteristic parameters from the source data pool; Based on the extracted core degradation feature parameters, a three-level degradation inflection point determination rule is set, and evolutionary law analysis is performed. The inflection point calibration results and evolutionary pattern analysis conclusions are combined with the original detection data from the source data pool to generate a standardized analysis report.

7. The method for analyzing changes in potato storage quality according to claim 6, characterized in that, Evolutionary pattern analysis specifically includes: temporal dimension analysis, spatial dimension analysis, and environmental coupling analysis.

8. A potato storage quality change analysis system, characterized in that, include: The benchmark database construction module is used to assign a unique identifier to each potato sample to be analyzed and to build its own anchor benchmark database. The homogeneous time series dataset acquisition module is used to perform continuous, non-destructive hyperspectral imaging detection on potato samples with unified identification at preset fixed time intervals, and simultaneously collect the real-time storage environment parameter set at the corresponding time node to generate a homogeneous time series dataset with storage duration as the time axis. The homogeneous matching module is used to perform a spatiotemporal anchoring homogeneous matching algorithm based on the inherent anchor benchmark database and the anchor homogeneous time series dataset of the analysis samples, and extract the specific degradation time series difference and degradation spatial diffusion gradient time series sequence of the degradation point. The degradation surface solution module is used to solve and generate the full-cycle spatiotemporal coupled degradation surface of potatoes from local to global real degradation process based on the specific degradation time difference and degradation spatial diffusion gradient time sequence of potato degradation points. The analysis report output module is used to output an analysis report on the spatiotemporal evolution of storage quality of the corresponding potato sample based on the generated full-cycle spatiotemporal coupled deterioration surface.

9. A computer storage medium, characterized in that, include: At least one memory and at least one processor; Memory, used to store one or more program instructions; A processor for running one or more program instructions to execute a method for analyzing changes in potato storage quality as described in any one of claims 1-7.

Images