Data-driven ice cream cold chain transportation visualization method and system

Abstract

The application discloses an ice cream cold chain transportation visualization method and system based on data driving, relates to the technical field of cold chain logistics monitoring, and constructs a multi-source correlation index by collecting multi-source data, detects abnormal temperature based on a sliding window, and completes missing data; a temperature and time double-factor damage cumulative function is constructed, a door opening transient temperature rise correction coefficient and an environmental heat invasion decay function are introduced, a quality heat damage cumulative index is calculated, a future time period one-time correction damage increment is output based on a time series prediction network, a refrigeration performance decay curve is used to predict a cargo compartment temperature, a final correction damage increment is calculated, and a time series risk level is generated; the time series risk level is mapped to a geographic road network node, a dynamic risk thermodynamic map layer is generated, and whole-process visualization monitoring is carried out; the problems of difficult quantification of quality damage, lagging risk prediction and non-intuitive monitoring in ice cream cold chain transportation are effectively solved, and the risk prediction accuracy and the real-time performance and interpretability of the visualization monitoring are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of cold chain logistics monitoring technology, specifically to a data-driven visualization method and system for ice cream cold chain transportation. Background Technology

[0002] Ice cream is extremely sensitive to temperature. Even slight overheating during transportation can cause ice crystal recrystallization and fat separation, leading to irreversible thermal damage and quality deterioration. This thermal damage has a cumulative effect and can be quantified by a thermodynamic model based on time and temperature integrals. However, in actual transportation, the risk changes dynamically due to the variable environment. As cold chain delivery networks expand and the number of vehicles en route continues to increase, monitoring centers need to manage a large number of vehicles simultaneously. However, current visualization methods are mainly based on real-time temperature threshold alarms, displaying historical status through dashboards, trajectory points, or heat maps, lacking forward-looking predictions of thermal damage risk levels in future periods. A few systems attempt to extrapolate temperature trends, but fail to organically integrate data-driven predictive thermal damage risk levels with dynamic geospatial interfaces. Risk representation is not linked to the spatiotemporal location of vehicles en route, making it impossible for monitoring personnel to intuitively identify which vehicles are about to enter a high-risk state and when and where the risk will occur on electronic maps. This passive response mechanism prevents monitoring parties from locating and intervening in vehicles that are about to deteriorate in advance, losing opportunities for proactive intervention such as dispatching, transferring, or adjusting routes, resulting in frequent cargo damage and complaints.

[0003] To address the aforementioned shortcomings, a technical solution is provided. Summary of the Invention

[0004] The purpose of this invention is to solve the problems of difficulty in quantifying quality damage, lag in risk prediction, and lack of intuitive monitoring in the cold chain transportation of ice cream, and to propose a data-driven visualization method and system for the cold chain transportation of ice cream.

[0005] The objective of this invention can be achieved through the following technical solutions: A data-driven visualization method for ice cream cold chain transportation includes: S1: Acquire raw data from vehicle-mounted temperature control equipment, Beidou and GPS dual-mode positioning unit, cargo door magnetic sensor, and meteorological data along the route; construct a multi-source association index; detect abnormal temperature data through a sliding window and complete missing data to generate a cold chain status dataset. S2: Construct a two-factor damage accumulation function based on temperature and time, introduce the correction coefficient for transient temperature rise when the cargo compartment is opened and the environmental heat intrusion attenuation function, calculate the cumulative index of heat damage to ice cream quality, and generate a sequence of damage increments corrected for future periods. S3: Obtain the predicted future ambient temperature values ​​of the vehicle's current location and planned path, combine them with the cooling performance decay curve to predict the temperature fluctuation range of the cargo compartment, correct the first-order correction damage increment sequence generated by S2, and generate a time-series risk level sequence. S4: Map the risk level to the geographic road network nodes, use inverse distance weighted interpolation to diffuse the risk value to the surface region, introduce the road congestion coefficient and altitude gradient to correct the spatial attenuation parameter, and generate a risk heat map layer. S5: Overlay real-time vehicle location, historical trajectory gradient lines, future path prediction zones, and risk heat map layers on the geographic information platform base map, trigger graded visual early warning symbols based on risk levels, display cold chain parameter panels, and provide predictive visualization monitoring of the entire process.

[0006] Furthermore, the specific operation steps of S1 are as follows: Acquire raw data from vehicle-mounted temperature control equipment, Beidou and GPS dual-mode positioning unit, cargo door magnetic sensor, and meteorological data along the route; using the temperature data acquisition timestamp as a unified reference axis, employ the nearest neighbor time matching algorithm to associate GPS trajectory, door opening event, and meteorological data with the corresponding timestamp, and construct a multi-source data association table with a unified time index; Set a sliding time window, calculate the standard deviation of the temperature dataset within the window, and if the standard deviation is greater than the preset threshold and no cargo compartment opening event is detected in the same period, then abnormal data is identified and marked. For missing and abnormal data segments, several sets of temperature data points are extracted before and after the missing interval, and the data is completed using a piecewise cubic Hermite interpolation formula. Iterate through the full temperature time series corresponding to all temperature sensors in the cargo compartment to generate a cold chain status dataset.

[0007] Furthermore, the specific operation steps of S2 include: Based on the ice cream quality safety threshold temperature and the activation energy of quality deterioration, a two-factor damage rate function of temperature and time is constructed. The complete transportation process is divided into several continuous time segments. The average temperature of the cargo compartment in each segment is calculated. The basic damage increment of each segment is calculated by combining the temperature and time two-factor damage rate function. Extract the magnetic sensor data of the cargo compartment door in each time segment and calculate the correction coefficient for transient temperature rise when the cargo compartment door is opened; Construct an environmental heat intrusion attenuation function and calculate the corresponding environmental heat intrusion correction coefficient; The basic damage increment is corrected based on the transient temperature rise correction coefficient when the cargo compartment is opened and the environmental heat intrusion correction coefficient, so as to obtain the first-corrected damage increment for each time segment. The first-corrected damage increments of all time segments are accumulated to obtain the current cumulative thermal damage index of ice cream quality.

[0008] Furthermore, the specific operation steps of S2 also include: Historical cold chain transportation data is retrieved, and the historical data is segmented into driving segments based on the road network segmentation standard. The ratio of the first-time damage increment to the cumulative duration is extracted for all time segments within a single road segment to obtain the historical damage rate gradient of the corresponding road segment. The historical damage rate gradient is used as a training label to train the temporal prediction network; The system acquires the attribute features of each driving segment in the planned route, constructs and trains a long short-term memory network for time-series prediction, and takes the segment attribute features, the current ice cream quality thermal damage accumulation index, and the real-time temperature status data of the entire cargo compartment as inputs. It outputs the first-time correction damage increment corresponding to each consecutive time period and generates a sequence of first-time correction damage increments for future time periods.

[0009] Furthermore, the specific operation steps of S3 include: Obtain the vehicle's current geographical location and the estimated arrival time of each future driving path node in the planned route; call the meteorological service interface to extract the predicted ambient temperature value of each node at the corresponding time and the corresponding prediction error fluctuation range; construct the cooling performance decay curve and calculate the actual effective cooling power of the equipment after cumulative operation. Extract the vehicle speed corresponding to the future path node, and predict the internal temperature of the cargo compartment by combining it with the actual effective cooling power. Based on the synchronous solution of the environmental temperature prediction error range, the temperature fluctuation range of the cargo compartment in multiple future time periods is obtained. Substitute the predicted cargo compartment temperature into the temperature and time dual-factor damage rate function, and combine the driving period duration to calculate the predicted baseline damage increment. Obtain the first corrected damage increment for the corresponding time period from the S2 output, and perform weighted fusion with the predicted baseline damage increment to obtain the final corrected damage increment for each future time period. By summing the current cumulative thermal damage index of ice cream quality with the final corrected damage increment for each future time period, the predicted cumulative damage value for each path node in the future is obtained.

[0010] Furthermore, the specific operation steps of S3 also include: A risk level determination threshold sequence is constructed, which is divided into five levels: normal, low risk, medium risk, high risk and extremely high risk. Based on the numerical range of the cumulative damage prediction value of each future path node, the corresponding risk level label is matched for each future time period. Based on the temperature fluctuation range of the cargo compartment, the cumulative damage prediction values ​​under the optimal and extreme working conditions are calculated respectively, and the corresponding risk levels are determined respectively; Based on the difference in risk levels between optimal and extreme operating conditions, different confidence levels are assigned to the corresponding risk levels. Traverse all future path nodes, integrate the expected arrival time, risk level and corresponding confidence level of each node, generate a time series risk level sequence, and mark the corresponding time and geographical coordinates of the first occurrence of medium risk or above in the sequence.

[0011] Furthermore, the specific operation steps of S4 are as follows: Extract all geographic road network nodes within the coverage area of ​​the vehicle planning path, and use the nearest neighbor matching algorithm to map the driving path nodes corresponding to the time-series risk level sequence to the geographic road network nodes one by one, and assign the corresponding final corrected damage increment as the original risk value to each road network node. Using the road network nodes that have completed risk assignment as the baseline data source, the risk values ​​of surrounding blank nodes are calculated through an inverse distance weighted interpolation algorithm. Obtain the congestion coefficient of each road network node and the corresponding road segment, and correct the original risk value based on the congestion coefficient; Collect elevation data of each road network node, calculate the elevation gradient between adjacent nodes, and set an elevation correction coefficient for uphill road sections where the elevation gradient exceeds a preset threshold. The final risk value of the road network node is calculated based on the corrected risk value and the altitude correction factor. The final risk values ​​of all road network nodes are filled into the pixel grid of the geographic base map to generate risk area distribution data; Construct a mapping relationship between risk level and color; generate single-moment risk heat map layers based on the timestamps bound to each node, and summarize all time-series layers to obtain a heat map sequence.

[0012] Furthermore, the specific operation steps of S5 are as follows: Based on the geographic information platform base map, real-time vehicle location icons, historical trajectory gradient lines, future path prediction bands, and risk heat map layers are superimposed and rendered from bottom to top. The vehicle location icon is refreshed based on vehicle positioning data, historical trajectory segments are generated by collecting positioning points over continuous time periods, and trajectory color gradient rules are set based on the final correction of damage increment. A semi-transparent future path prediction zone is generated along the boundary of the planned route; the risk level judgment threshold is retrieved, the risk value of each location in the heat map is compared, and graded visual warning symbols are attached to the corresponding points and refreshed with the heat map layer in time. Establish an interactive monitoring mechanism, which, upon triggering, retrieves all cold chain data of the target vehicle, generates a parameter panel, and displays it in a pop-up window; Calculate the overall warning index of the screen, and trigger global screen border flashing when the overall warning index of the screen is greater than the preset threshold; The display range of each element is adjusted based on the layer linkage mechanism, and the risk distribution at different prediction times is switched based on the time axis control.

[0013] A second aspect of the present invention provides a data-driven visualization system for ice cream cold chain transportation, comprising: Data preprocessing module: Collects multi-source data to build a multi-source association index; filters out abnormal data based on a sliding window and performs interpolation to complete the data, outputting a cold chain status dataset; Damage Calculation and Training Module: Calculates the basic damage increment, introduces the transient temperature rise correction coefficient for cargo compartment opening and the environmental thermal intrusion attenuation function for correction, and generates a first-time corrected damage increment sequence for future periods; Damage fusion and classification module: Constructs a cooling performance degradation curve to predict the temperature fluctuation range of the cargo compartment, corrects the first-order damage increment sequence, and generates a time-series risk level sequence; Heatmap generation module: The node risk level is diffused through inverse distance weighted interpolation, and corrected based on the congestion coefficient and altitude gradient to generate a time-series dynamic risk heatmap layer; Visualization module: Overlays vehicle locations, historical trajectory gradient lines, future path prediction zones, and risk heat map layers on the geographic information platform base map, displays cold chain parameters in tiered early warning pop-ups, and performs full-process visual monitoring.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention constructs a dual-factor damage accumulation function based on temperature and time, introduces corrections for transient temperature rise upon door opening and attenuation of environmental thermal intrusion, combines an LSTM network to predict future damage increments, and dynamically corrects risk levels based on refrigeration performance degradation and vehicle speed-induced wind cooling effects. Simultaneously, it maps risks to geographical road network nodes, generates a dynamic risk heat map layer after correction by congestion coefficients and altitude gradients, and overlays vehicle trajectories and graded warning symbols to achieve full-process visualized monitoring. This significantly improves the quantitative accuracy of quality damage and the forward-looking nature of risk prediction in ice cream cold chain transportation, and enhances the intuitiveness and real-time interactivity of the monitoring interface. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0016] 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.

[0017] Example: like Figure 1 As shown, a data-driven visualization method for ice cream cold chain transportation includes: S1: Collect data from vehicle temperature control, GPS trajectory, door opening sensors, and weather along the route; construct a data association index based on timestamp alignment; identify and mark anomalies through a sliding time window, complete missing data, and output a spatiotemporally continuous cold chain status dataset. The specific implementation process is as follows: Acquire raw data from the vehicle-mounted temperature control equipment, the Beidou and GPS dual-mode positioning unit, and the magnetic sensor on the cargo door, and call the external meteorological service interface to retrieve meteorological data along the route; The on-board temperature control equipment uses a communication interface to collect real-time temperature data from temperature sensors at various points inside the cargo compartment based on a preset collection frequency, and constructs a temperature time series. The system continuously collects geographic coordinate information during vehicle travel using a dual-mode positioning unit combining BeiDou and GPS, generates a latitude and longitude coordinate sequence, and calculates the vehicle's instantaneous speed. The opening and closing actions of the cargo door are captured in real time by the magnetic sensor of the cargo door, forming a door opening event stream, including the timestamps of the opening, closing, and opening / closing times, as well as the duration of a single door opening action; Call the meteorological service interface and obtain the ambient temperature along the route by matching point by point based on the real-time trajectory coordinates of the vehicle. Using the temperature data collection timestamp as a unified reference time axis, and employing the nearest neighbor time matching algorithm, GPS trajectory coordinates, cargo compartment door opening events, and various types of meteorological data are associated with their corresponding timestamps to construct a multi-source data association table with a unified time index. Set a sliding time window, extract temperature data within the sliding window range, and calculate the standard deviation of the temperature data. When the standard deviation of the temperature data within the sliding window range is greater than a preset threshold, and no cargo compartment opening event is detected in the same period, the data in that period is determined to be abnormal data and marked. For data gaps caused by communication interruptions and marked abnormal data segments, extract several sets of temperature data points before and after the missing interval, and then use the formula... Data interpolation is performed to complete missing or abnormal data segments; among which, Indicates the time after interpolation calculation The corresponding repair temperature value, This represents the original timestamp corresponding to the data point to be repaired. Describes the piecewise cubic Hermitian basis functions. The basis functions correspond to the function values ​​at the left and right endpoints, respectively. The derivative basis functions corresponding to the left and right endpoints, respectively. , representing the normalized time parameter, These are the timestamps of the adjacent valid data points on the left and right sides of the missing or abnormal data segment, respectively. Representing the left boundary points respectively Right boundary point The slope of the tangent line to the temperature time series at that point. This represents a fixed time interval between the left and right boundary points; The system iterates through the full temperature time series corresponding to all temperature sensors in the cargo compartment, sequentially performing anomaly detection, anomaly marking, and missing data interpolation repair to generate a spatiotemporally continuous, multi-dimensional information-associative cold chain status dataset.

[0018] S2: Construct a two-factor damage accumulation function based on temperature and time, introduce a correction coefficient for transient temperature rise when the cargo compartment door is opened and an environmental heat intrusion attenuation function, calculate the cumulative thermal damage index of ice cream quality for each time period, and train a time-series prediction network based on the damage rate gradient of historical transportation data for the same route. The specific implementation process is as follows: Pre-determined safety threshold temperature for ice cream quality and activation energy for ice cream quality deterioration A two-factor damage rate function based on temperature and time is constructed, and its specific expression is as follows: ,in, Indicates the current temperature of the cargo compartment. Below is the rate of quality deterioration of ice cream per unit time, i.e., the rate of damage. The formula-related frequency factor was determined through data regression calibration using isothermal accelerated aging experiments combined with Arrhenius plots for high-fat cream ice cream. The range of values ​​is For regular hard ice cream The value range is 6.0. Low-fat popsicles The value range is 8.0. ; Represents an exponential function. Universal gas constant, This indicates the conversion of Celsius temperature to thermodynamic temperature; The complete transportation process is divided into several consecutive independent time segments, and the average temperature of the cargo compartment within each time segment is calculated and denoted as . Combining the temperature and time-based damage rate function, the incremental damage to the ice cream during this period is calculated using the following formula: ,in, Indicates the first The base damage increment for each time segment Indicates the first The duration of each time segment; Extract the magnetic sensing time data of the cargo compartment door within each time segment. If a door opening event is detected within a time segment, record the duration of that door opening event. Extract the ambient temperature corresponding to the moment the door opening event occurred. With cargo compartment temperature Through formula Calculations are performed to obtain the first Correction factor for transient temperature rise when cargo compartment doors are opened at a given time interval ,in, This indicates the impact of door opening on the base coefficient, with a value ranging from 0.3 to 0.8. Indicates the temperature difference sensitivity coefficient. Indicates the first The set of sequence numbers of all door-opening events that occur within a time segment. Indicates the first The external temperature of the cargo compartment during the second door opening incident. No. The internal temperature of the cargo compartment corresponding to the second door opening event; if no door opening event is detected in a certain time segment, then ; To indicate unit conversion, Unit conversion to minutes, unified dimension matching coefficient ; Based on the ambient temperature and cargo compartment temperature corresponding to each time segment, combined with the cargo compartment insulation performance coefficient... Effective heat transfer surface area of ​​the cargo compartment Ice cream specific heat capacity under constant pressure Total weight of goods Preset reference temperature rise Construct an environmental heat intrusion attenuation function, the specific expression of which is: ,in, Indicates the first Environmental thermal intrusion correction factor for each time segment Indicates the first The average ambient temperature along the route within a time segment; The thermal insulation performance coefficient of the cargo compartment is based on the testing standard for the thermal insulation performance of refrigerated cargo compartments and is obtained by conducting on-site thermal tests on cargo compartments in use. The effective heat transfer surface area of ​​the cargo compartment refers to the total area of ​​the cargo compartment that undergoes heat exchange with the outside air. Based on the transient temperature rise correction coefficient when the cargo compartment door opens and the environmental thermal intrusion attenuation function, the basic damage increment is corrected to obtain the first-order corrected damage increment corresponding to a single time segment. The revised calculation formula is as follows: ; By summing up the incremental damage from one correction across all time segments, we obtain the cumulative thermal damage index for the current ice cream quality. The calculation formula is: ,in, This represents the total number of all time segments up to the current moment; Historical cold chain transportation data with the same origin and destination and the same transportation season are retrieved. Based on the road network segmentation standard matched for this transportation, the historical data is divided into driving segments. The ratio of the corrected damage increment under all time segments in a single road segment to the cumulative duration of all time segments in a single road segment is calculated to obtain the historical damage rate gradient corresponding to each road segment, which is used to reflect the average rate of deterioration of ice cream quality in that road segment. The historical damage rate gradient is used as a training label to train the temporal prediction network; The road network segmentation standard is a rule for uniformly delineating road segment boundaries based on geographical road network nodes, road intersections, administrative divisions, and mileage thresholds; Obtain the attribute features of each driving segment in the planned path, including: The road attributes and road condition attributes of each road segment are collected as road segment features. The date, time period, and solar term information of historical transportation tasks are collected as time features. The average cargo compartment temperature, temperature fluctuation amplitude, and average ambient temperature of the entire road segment are collected as temperature features. The average cargo compartment temperature is the arithmetic mean of the cargo compartment temperature at all collection times within a single road segment, and the average ambient temperature is the arithmetic mean of the ambient temperature along the route within a single road segment. A long short-term memory (LSTM) network for time-series prediction was constructed and trained. This network takes road segment attribute features, time features, temperature features, and the current moment's cumulative thermal damage index of ice cream quality and real-time temperature data of the entire cargo compartment as inputs, and uses the historical damage rate gradient of the corresponding road segment as the output label for training. During prediction, the current moment's cumulative thermal damage index of ice cream quality, real-time temperature data of the entire cargo compartment, and attribute features of each future travel segment in the planned route are input into the trained network, which outputs the predicted damage rate gradient for each future road segment. The predicted damage rate gradient is multiplied by the corresponding future travel time period to obtain the first-order corrected damage increment for each consecutive time period, generating a complete sequence of first-order corrected damage increments for the future time period.

[0019] S3: Obtain the predicted ambient temperature for the future time period based on the vehicle's current location and planned path. Combine this with the vehicle's historical cooling performance degradation curve to predict temperature fluctuation ranges for multiple time periods. Dynamically correct the future time period modified damage increment sequence generated in S2 to generate the final time-series risk level sequence and confidence level. The specific implementation process is as follows: Obtain the latitude and longitude coordinates of the vehicle's current geographical location, and at the same time obtain the time sequence of the vehicle's future driving path nodes output by the navigation plan, including the geographical location latitude and longitude information of each path node and the estimated arrival time of the vehicle at the corresponding node. Call the cloud-based weather forecast service interface to extract the predicted ambient temperature values ​​of each future driving path node at the corresponding expected arrival time in batches, and simultaneously match and obtain the ambient temperature prediction fluctuation error range at the corresponding time and geographical location, and define the minimum and maximum predicted ambient temperature values. Retrieve the refrigeration equipment's operating parameters from the vehicle's historical transportation data, including rated refrigeration power, cumulative operating time, and maintenance records, and construct a refrigeration performance degradation curve, expressed as: ,in, Indicates the cumulative operation of refrigeration equipment The actual effective cooling capacity after that, This indicates the factory-specified rated cooling capacity of the refrigeration unit. This represents the performance degradation rate coefficient of refrigeration equipment, with a value range of 0.0001-0.0005. This indicates the rated design service life of the entire refrigeration equipment as specified by the manufacturer. Extract the vehicle speed corresponding to the future path nodes, and combine it with the current temperature of the cargo compartment, the specific heat capacity of ice cream at constant pressure, and the total mass of the cargo to calculate the predicted cargo compartment temperature for that period. The calculation formula is as follows: ,in, Indicates the first Predicted internal temperature of the cargo compartment for each path node. This indicates the current measured temperature inside the cargo compartment of the vehicle. Indicates the first Predicted ambient temperature for each path node. Indicates reaching the th The duration of the travel time corresponding to each path node This represents the thermal time constant of the cargo compartment, characterizing how quickly the compartment responds to temperature changes due to ambient temperature differences. It is calibrated by actual measurements of the cargo compartment's insulation structure and volume parameters, and the calculation formula is: , This indicates the overall equivalent heat capacity of the cargo compartment. This indicates the overall heat transfer coefficient of the cargo compartment enclosure structure. This indicates the total heat exchange surface area of ​​the cargo compartment's outer wall; Indicates the actual effective cooling capacity. Indicates driving speed The corresponding wind-cooling effect coefficient indicates that the higher the vehicle's wind speed, the stronger the wind-cooling effect, reflecting the enhancing effect of vehicle speed on the cooling system's heat dissipation efficiency. Based on the environmental temperature prediction error range, the predicted cargo compartment temperature is simultaneously substituted into the calculation formula to obtain the upper and lower bounds of the cargo compartment temperature fluctuation, and thus the cargo compartment temperature fluctuation range for multiple future time periods. Substituting the predicted cargo compartment temperature into the temperature-time dual-factor damage rate function, and combining it with the duration of the travel period, the formula is used. Calculations are performed to obtain the first Predicted baseline damage increment for each path node; Based on the corrected damage increment corresponding to this time period output by S2, the formula is used. The final corrected damage increment is obtained through calculation, where, This indicates the corrected damage increment for the corresponding time period in the S2 output. The weighting coefficients representing the corresponding time period corrections for damage increments in the S2 output range from 0 to 1 and are obtained through fitting and calibration based on historical transport samples. This represents the probability of door opening based on historical door opening frequency statistics. It is calculated using the S2 door opening correction coefficient formula, taking into account the duration of a single door opening, environmental conditions, and predicted cargo compartment temperature, to obtain the transient temperature rise correction coefficient for cargo compartment door opening. This is useful if historical door opening statistics for a road segment are unavailable or the door opening probability cannot be quantified. The value is assigned to 1; This represents the environmental heat intrusion correction coefficient calculated using the S2 environmental heat intrusion correction coefficient calculation formula, based on the predicted ambient temperature and the predicted cargo compartment temperature. By summing the current cumulative thermal damage index of ice cream quality with the final corrected damage increment for each future time period, the cumulative damage prediction value corresponding to each path node in the future is obtained. Four sets of thresholds are preset to form a risk level determination threshold sequence, which are divided into five levels: normal, low risk, medium risk, high risk and extremely high risk. Based on the numerical range of the cumulative damage prediction value, the corresponding risk level label is matched for each future time period. Based on the temperature fluctuation range of the cargo compartment, the cumulative damage prediction value corresponding to the extreme value of the range is calculated, that is, the cumulative damage prediction value under the optimal working condition and the extreme working condition. Based on the risk level determination threshold sequence, the risk level corresponding to the cumulative damage prediction value under the optimal working condition and the extreme working condition is determined respectively; If the risk levels corresponding to the cumulative damage prediction values ​​of the optimal working condition and the extreme working condition are the same, the confidence level of the corresponding risk level is assigned to 1. If the risk levels corresponding to the cumulative damage prediction values ​​of the optimal working condition and the extreme working condition differ by one level, the confidence level of the corresponding risk level is assigned a value of 0.7. If the risk levels corresponding to the cumulative damage prediction values ​​of the optimal working condition and the extreme working condition differ by two levels or more, the confidence level of the corresponding risk level is assigned a value of 0.4. Traverse all future path nodes, integrate the expected arrival time, risk level and corresponding confidence level of each node, generate a time series risk level sequence, and mark the corresponding time and geographical coordinates of the first occurrence of medium risk or above in the sequence.

[0020] S4: Map the temporal risk level to geographic road network nodes, use inverse distance weighted interpolation to diffuse the risk value to the surface region, introduce road congestion coefficient and elevation gradient to correct spatial attenuation parameters, and dynamically generate a risk heat map layer with a time dimension. The specific implementation process is as follows: The road network topology structure stored in the geographic information platform is taken, and all road network nodes within the coverage area of ​​the vehicle planning path are extracted to construct a set of road network nodes. The nearest neighbor matching algorithm is used to match and map the driving path nodes corresponding to the predicted time-series risk level sequence to the geographic road network nodes one by one. Each road network node is assigned the corresponding single-time period final corrected damage increment as the original risk value and bound to the timestamp of the corresponding time, thus completing the association and binding of road network node risk data and time-series data. Using the road network nodes that have completed risk assignment as the baseline data source, the risk values ​​of surrounding blank nodes are calculated through an inverse distance weighted interpolation algorithm. The calculation formula is: ,in, This represents the complete set of baseline road network nodes that participated in the interpolation calculation and have been labeled with their original risks. This represents the original risk value inherent to the baseline road network nodes. This represents the geographical distance between the node to be interpolated and a single baseline road network node. This represents the distance decay index, with a value range of 2.0-3.0; Access the traffic information service interface to obtain the congestion coefficient of each road network node and corresponding road segment. The original risk value is corrected based on the congestion coefficient, and the calculation formula is as follows: ,in, This represents the risk value of road network nodes after correction for road congestion factors. This represents the original risk value of a road network node before congestion correction. Collect elevation data for each road network node and calculate the elevation gradient between adjacent nodes. For altitude gradients exceeding a preset threshold Altitude correction factor is set for uphill sections. : ; Based on the corrected road network node risk values ​​and altitude correction coefficients, using the formula... The final risk value of the road network nodes is calculated. The final risk values ​​of all road network nodes are rasterized and filled into the geographic base map pixel grid. A Gaussian kernel function is used to perform spatial smoothing on the raster data to generate continuous risk area distribution data. Construct a risk level and color mapping relationship: normal corresponds to dark green, low risk corresponds to light green, medium risk corresponds to yellow, high risk corresponds to orange, and extremely high risk corresponds to red. Generate gradient colors based on linear interpolation of risk values. Based on the timestamps bound to each node, a single-moment risk heatmap layer is generated. All time-series layers are then aggregated to obtain a heatmap sequence with a time dimension. The heatmap sequence supports dynamic playback along the time axis and querying of slices at specified times.

[0021] S5: Overlay real-time vehicle location icons, historical trajectory gradient lines, future path prediction bands, and risk heat map layers onto the geographic information platform base map. Trigger graded visual warning symbols based on risk level thresholds, and display vehicle cold chain parameter panels in response to user interaction operations. This achieves predictive and visual monitoring of the entire ice cream cold chain transportation process. The specific implementation process is as follows: Based on the geographic information platform base map as the basic layer, four types of visualization elements are loaded in layers, including real-time vehicle location icons, historical trajectory gradient lines, future path prediction strips, and risk heat map layers. All types of layers are superimposed and rendered in a hierarchical order from bottom to top. The vehicle location icon is refreshed based on the location data periodically transmitted by the vehicle positioning device. The icon is bound to the vehicle's unique code and the current time. The vehicle's location points are collected for continuous periods of time and arranged in the driving sequence to generate historical trajectory segments. The trajectory color gradient rule is set based on the final corrected damage increment value corresponding to each point along the way. As the damage increment increases, the trajectory display color tone is switched synchronously. Based on the vector data of the planned driving routes, a future path prediction band with a semi-transparent fill pattern is generated along the route boundary, and the range of the prediction band matches the actual direction of the road network. The system retrieves preset risk level thresholds and compares them with the final risk values ​​at each location within the heat map. Based on the comparison results, it matches graded visual warning symbols. When the risk value enters the corresponding threshold range, a dedicated warning icon is attached to the corresponding geographical location. The warning icon is dynamically refreshed in sync with the underlying heat map layer. A front-end interactive listening mechanism is set up. After the interactive action is triggered, the full amount of cold chain data collected from the target vehicle is retrieved. The data includes the real-time temperature of the cargo compartment, door opening records, real-time ambient temperature, and current heat damage cumulative index. The detailed cold chain parameter panel of the vehicle is generated according to the preset layout rules and displayed in a pop-up window. The comprehensive early warning index for the visual display is set, and the calculation formula is as follows: ,in, This indicates the overall warning index of the image. This indicates the number of high-risk and extremely high-risk locations within the current field of view. This indicates the number of medium-risk locations currently within the field of view. This indicates the weighting coefficients for high-risk and extremely high-risk locations. This represents the weighting coefficient for medium-risk locations. ; When the overall warning index of the screen exceeds the preset threshold, an auxiliary warning effect of flashing the global screen border is triggered simultaneously; Based on the layer linkage mechanism, multi-element data linkage is completed. When zooming and panning the base map, the display range of vehicle icons, trajectory lines, path prediction bands, heat maps, and warning symbols are adjusted synchronously. Based on the time axis control of the time series heat map, the global risk distribution at different prediction times is switched, and predictive visualization monitoring of the entire ice cream cold chain transportation process is completed.

[0022] A data-driven visualization system for ice cream cold chain transportation includes: Data preprocessing module: Collects vehicle temperature control, GPS trajectory, door opening sensor and weather data along the route, builds data association index based on timestamp alignment; identifies and marks anomalies through sliding time windows, completes the data for anomalies, and generates cold chain status dataset; Damage Calculation and Training Module: Construct a dual-factor damage accumulation function based on temperature and time, introduce a correction coefficient for transient temperature rise when the cargo compartment is opened and an environmental heat intrusion attenuation function, calculate the cumulative thermal damage index of ice cream quality for each time period, and train a time-series prediction network based on the damage rate gradient of historical transportation data on the same route, and output a first-corrected damage increment sequence. Damage fusion and classification module: Obtain the predicted ambient temperature values ​​for future time periods based on the vehicle's current location and planned path, combine them with the vehicle's historical cooling performance decay curve, predict temperature fluctuation ranges for multiple time periods, dynamically correct the generated first-order modified damage increment sequence, and generate a time-series risk level sequence and corresponding confidence level. Heatmap generation module: Maps time-series risk levels to geographic road network nodes, uses inverse distance weighted interpolation to diffuse risk values ​​to the surface region, introduces road congestion coefficient and altitude gradient to correct spatial attenuation parameters, and generates risk heatmap layer; Visualization module: Real-time vehicle location icons, historical trajectory gradient lines, future path prediction bands, and risk heat map layers are overlaid on the geographic information platform base map. Based on the risk level threshold, graded visual early warning symbols are triggered to achieve predictive visual monitoring of ice cream cold chain transportation.

[0023] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A data-driven visualization method for ice cream cold chain transportation, characterized in that, include: S1: Acquire raw data from vehicle-mounted temperature control equipment, Beidou and GPS dual-mode positioning unit, cargo door magnetic sensor, and meteorological data along the route; construct a multi-source association index; detect abnormal temperature data through a sliding window and complete missing data to generate a cold chain status dataset. S2: Construct a two-factor damage accumulation function based on temperature and time, introduce the correction coefficient for transient temperature rise when the cargo compartment is opened and the environmental heat intrusion attenuation function, calculate the cumulative index of heat damage to ice cream quality, and generate a sequence of damage increments corrected for future periods. S3: Obtain the predicted future ambient temperature values ​​of the vehicle's current location and planned path, combine them with the cooling performance decay curve to predict the temperature fluctuation range of the cargo compartment, correct the first-order correction damage increment sequence generated by S2, and generate a time-series risk level sequence. S4: Map the risk level to the geographic road network nodes, use inverse distance weighted interpolation to diffuse the risk value to the surface region, introduce the road congestion coefficient and altitude gradient to correct the spatial attenuation parameter, and generate a risk heat map layer. S5: Overlay real-time vehicle location, historical trajectory gradient lines, future path prediction zones, and risk heat map layers on the geographic information platform base map, trigger graded visual early warning symbols based on risk levels, display cold chain parameter panels, and provide predictive visualization monitoring of the entire process.

2. The data-driven visualization method for ice cream cold chain transportation according to claim 1, characterized in that, The specific operation steps of S1 are as follows: Acquire raw data from vehicle-mounted temperature control equipment, Beidou and GPS dual-mode positioning unit, cargo door magnetic sensor, and meteorological data along the route; using the temperature data acquisition timestamp as a unified reference axis, employ the nearest neighbor time matching algorithm to associate GPS trajectory, door opening event, and meteorological data with the corresponding timestamp, and construct a multi-source data association table with a unified time index; Set a sliding time window, calculate the standard deviation of the temperature dataset within the window, and if the standard deviation is greater than a preset threshold and no cargo compartment opening event is detected in the same period, then abnormal data is identified and marked. For missing and abnormal data segments, several sets of temperature data points are extracted before and after the missing interval, and the data is completed using a piecewise cubic Hermite interpolation formula. Iterate through the full temperature time series corresponding to all temperature sensors in the cargo compartment to generate a cold chain status dataset.

3. The data-driven visualization method for ice cream cold chain transportation according to claim 1, characterized in that, The specific operation steps of S2 include: Based on the ice cream quality safety threshold temperature and the activation energy of quality deterioration, a two-factor damage rate function of temperature and time is constructed. The complete transportation process is divided into several continuous time segments. The average temperature of the cargo compartment in each segment is calculated. The basic damage increment of each segment is calculated by combining the temperature and time two-factor damage rate function. Extract the magnetic sensor data of the cargo compartment door in each time segment and calculate the correction coefficient for transient temperature rise when the cargo compartment door is opened; Construct an environmental heat intrusion attenuation function and calculate the corresponding environmental heat intrusion correction coefficient; The basic damage increment is corrected based on the transient temperature rise correction coefficient when the cargo compartment is opened and the environmental heat intrusion correction coefficient, so as to obtain the first-corrected damage increment for each time segment. The first-corrected damage increments of all time segments are accumulated to obtain the current cumulative thermal damage index of ice cream quality.

4. The data-driven visualization method for ice cream cold chain transportation according to claim 3, characterized in that, The specific operation steps of S2 also include: Historical cold chain transportation data is retrieved, and the historical data is segmented into driving segments based on the road network segmentation standard. The ratio of the first-time damage increment to the cumulative duration is extracted for all time segments within a single road segment to obtain the historical damage rate gradient of the corresponding road segment. The historical damage rate gradient is used as a training label to train the temporal prediction network; The system acquires the attribute features of each driving segment in the planned route, constructs and trains a long short-term memory network for time-series prediction, and takes the segment attribute features, the current ice cream quality thermal damage accumulation index, and the real-time temperature status data of the entire cargo compartment as inputs. It outputs the first-time correction damage increment corresponding to each consecutive time period and generates a sequence of first-time correction damage increments for future time periods.

5. The data-driven visualization method for ice cream cold chain transportation according to claim 1, characterized in that, The specific operation steps of S3 include: Obtain the vehicle's current geographical location and the estimated arrival time of each future driving path node in the planned route; call the meteorological service interface to extract the predicted ambient temperature value of each node at the corresponding time and the corresponding prediction error fluctuation range; construct the cooling performance decay curve and calculate the actual effective cooling power of the equipment after cumulative operation. Extract the vehicle speed corresponding to the future path node, and predict the internal temperature of the cargo compartment by combining it with the actual effective cooling power. Based on the synchronous solution of the environmental temperature prediction error range, the temperature fluctuation range of the cargo compartment in multiple future time periods is obtained. Substitute the predicted cargo compartment temperature into the temperature and time dual-factor damage rate function, and combine the driving period duration to calculate the predicted baseline damage increment. Obtain the first corrected damage increment for the corresponding time period from the S2 output, and perform weighted fusion with the predicted baseline damage increment to obtain the final corrected damage increment for each future time period. By summing the current cumulative thermal damage index of ice cream quality with the final corrected damage increment for each future time period, the predicted cumulative damage value for each path node in the future is obtained.

6. The data-driven visualization method for ice cream cold chain transportation according to claim 5, characterized in that, The specific operation steps of S3 also include: A risk level determination threshold sequence is constructed, which is divided into five levels: normal, low risk, medium risk, high risk and extremely high risk. Based on the numerical range of the cumulative damage prediction value of each future path node, the corresponding risk level label is matched for each future time period. Based on the temperature fluctuation range of the cargo compartment, the cumulative damage prediction values ​​under the optimal and extreme working conditions are calculated respectively, and the corresponding risk levels are determined respectively; Based on the difference in risk levels between optimal and extreme operating conditions, different confidence levels are assigned to the corresponding risk levels. Traverse all future path nodes, integrate the expected arrival time, risk level and corresponding confidence level of each node, generate a time series risk level sequence, and mark the corresponding time and geographical coordinates of the first occurrence of medium risk or above in the sequence.

7. The data-driven visualization method for ice cream cold chain transportation according to claim 1, characterized in that, The specific operation steps of S4 are as follows: Extract all geographic road network nodes within the coverage area of ​​the vehicle planning path, and use the nearest neighbor matching algorithm to map the driving path nodes corresponding to the time-series risk level sequence to the geographic road network nodes one by one, and assign the corresponding final corrected damage increment as the original risk value to each road network node. Using the road network nodes that have completed risk assignment as the baseline data source, the risk values ​​of surrounding blank nodes are calculated through an inverse distance weighted interpolation algorithm. Obtain the congestion coefficient of each road network node and the corresponding road segment, and correct the original risk value based on the congestion coefficient; Collect elevation data of each road network node, calculate the elevation gradient between adjacent nodes, and set an elevation correction coefficient for uphill road sections where the elevation gradient exceeds a preset threshold. The final risk value of the road network node is calculated based on the corrected risk value and the altitude correction factor. The final risk values ​​of all road network nodes are filled into the pixel grid of the geographic base map to generate risk area distribution data; Establish a mapping relationship between risk levels and colors; Based on the timestamps bound to each node, a single-moment risk heat map layer is generated, and all time-series layers are aggregated to obtain a heat map sequence.

8. The data-driven visualization method for ice cream cold chain transportation according to claim 1, characterized in that, The specific operation steps of S5 are as follows: Based on the geographic information platform base map, real-time vehicle location icons, historical trajectory gradient lines, future path prediction bands, and risk heat map layers are superimposed and rendered from bottom to top. The vehicle location icon is refreshed based on vehicle positioning data, historical trajectory segments are generated by collecting positioning points over continuous time periods, and trajectory color gradient rules are set based on the final correction of damage increment. A semi-transparent future path prediction zone is generated along the boundary of the planned route. Retrieve the risk level judgment threshold, compare the risk value of each location in the heat map, attach graded visual warning symbols at the corresponding points and refresh them in sequence with the heat map layer; Establish an interactive monitoring mechanism, which, upon triggering, retrieves all cold chain data of the target vehicle, generates a parameter panel, and displays it in a pop-up window; Calculate the overall warning index of the screen, and trigger global screen border flashing when the overall warning index of the screen is greater than the preset threshold; The display range of each element is adjusted based on the layer linkage mechanism, and the risk distribution at different prediction times is switched based on the time axis control.

9. A system applied to the data-driven visualization method for ice cream cold chain transportation as described in any one of claims 1-8, comprising: Data preprocessing module: Collects data from multiple sources and constructs a multi-source correlation index; The cold chain status dataset is output by filtering out abnormal data using a sliding window and performing interpolation to complete the dataset. Damage Calculation and Training Module: Calculates the basic damage increment, introduces the transient temperature rise correction coefficient for cargo compartment opening and the environmental thermal intrusion attenuation function for correction, and generates a first-time corrected damage increment sequence for future periods; Damage fusion and classification module: Constructs a cooling performance degradation curve to predict the temperature fluctuation range of the cargo compartment, corrects the first-order damage increment sequence, and generates a time-series risk level sequence; Heatmap generation module: The node risk level is diffused through inverse distance weighted interpolation, and corrected based on the congestion coefficient and altitude gradient to generate a time-series dynamic risk heatmap layer; Visualization module: Overlays vehicle locations, historical trajectory gradient lines, future path prediction zones, and risk heat map layers on the geographic information platform base map, displays cold chain parameters in tiered early warning pop-ups, and performs full-process visual monitoring.

Images