A method and system for real-time monitoring and analysis of food color change

Abstract

The application provides a real-time monitoring and analysis method and system for color change of food. Wherein, the color change image of the food is captured, and a multi-modal food color dataset is constructed; based on the multi-modal food color dataset, dynamic correction processing is performed on the color change of the food, and a color change trend analysis report is obtained; according to the color change trend analysis report, the decay rate of the freshness of the food and the possible deterioration point are predicted, and a food shelf life evaluation result is generated; using the food shelf life evaluation result, an intelligent logistics scheduling system driven by a multi-objective optimization algorithm is used to automatically plan the whole process path of the food from production to consumption, dynamically adjust the storage conditions and the transportation time window, and generate a logistics scheme. The technical scheme provided by the application realizes more accurate dynamic correction processing of the color change of the food, thereby significantly improving the accuracy of the prediction of the freshness of the food.

Description

TECHNICAL FIELD

[0001] The embodiments of the present application relate to the technical field of real-time monitoring, and in particular to a real-time monitoring and analysis method and system for food color change. BACKGROUND

[0002] With the increasing demand of consumers for food safety and freshness, real-time monitoring and analysis technology in the food supply chain becomes particularly important.

[0003] Currently, some food industries use traditional visual detection systems combined with manual inspection to preliminarily assess food color change. These methods usually rely on fixed-angle photos or video streams and simple environmental sensor data collection, and then use basic data processing algorithms to identify potential problem areas.

[0004] Traditional methods are difficult to fully capture all details of food color change, especially when light source conditions change, which can easily lead to misjudgment; the prediction ability of food color change trend is limited, and it is difficult to accurately assess the freshness decay rate and possible deterioration points of food. SUMMARY

[0005] The embodiments of the present application provide a real-time monitoring and analysis method and system for food color change to solve the problem of limited prediction ability of food color change trend in the prior art.

[0006] In a first aspect, the embodiments of the present application provide a real-time monitoring and analysis method for food color change, comprising:

[0007] Real-time capture of food color change images using high-sensitivity imaging devices from multiple angles and different spectral ranges, collection of temperature and humidity, light intensity parameters of the environment where the food is located using environmental perception devices, and construction of a multi-modal food color dataset;

[0008] Based on the multi-modal food color dataset, dynamic correction processing of food color change is performed by combining a random forest algorithm under an ensemble learning framework with an adaptive color space mapping technology, and a color change trend analysis report is obtained;

[0009] According to the color change trend analysis report, the decay rate of food freshness and possible deterioration points are predicted by combining a long short-term memory network based on time series analysis with food composition analysis technology and a gray prediction model, and a food shelf life evaluation result is generated;

[0010] Using the food shelf life evaluation result, the full-process path of food from production to consumption is automatically planned and processed by an intelligent logistics scheduling system driven by a multi-objective optimization algorithm, and the storage conditions and transportation time window are dynamically adjusted to generate a logistics scheme.

[0011] Optionally, based on the multi-modal food color dataset, a random forest algorithm under an ensemble learning framework is combined with an adaptive color space mapping technique to dynamically correct the food color change, and a color change trend analysis report is obtained, including:

[0012] The multi-modal food color dataset is used to perform adaptive color space mapping processing on color information in different spectral ranges, eliminate color deviation caused by light source changes, and obtain standardized color data.

[0013] Based on the standardized color data, a random forest algorithm is combined with the color change pattern of food under different environmental conditions to perform prediction modeling processing on food color change, and a food color change prediction model is generated.

[0014] According to the food color change prediction model, trend and periodicity feature analysis processing is performed on the time series data of food color change, the color stability of food under different environmental conditions is evaluated, and a color change feature evaluation result is obtained.

[0015] Using the color change feature evaluation result, the main trend, periodicity feature, color change rate and amplitude information of food color change are integrated through data visualization technology and statistical analysis method, and a color change trend analysis report is generated.

[0016] Optionally, based on the standardized color data, a random forest algorithm is combined with the color change pattern of food under different environmental conditions to perform prediction modeling processing on food color change, and a food color change prediction model is generated, including:

[0017] Before calculating C pred (t), the multi-modal food color dataset needs to be preprocessed to eliminate color deviation caused by light source changes through adaptive color space mapping technology, and standardized color data is obtained to prepare for building a random forest model for predicting food color change.

[0018]

[0019] wherein C pred (t) represents the predicted food color change value at time t, w i is the weight of the i-th tree in the random forest, reflecting the contribution of each tree to the final prediction result; C i (t) represents the standardized color data of the i-th sample at time t; E i(t) represents the environmental condition data of the i-th sample at time t; N is the number of samples; a is an adjustment coefficient related to the environmental temperature, used to adjust the influence of temperature on color change; b is an adjustment coefficient related to the environmental humidity, used to adjust the influence of humidity on color change; T(t) represents the environmental temperature at time t; H(t) represents the environmental humidity at time t; d is an adjustment coefficient related to the ratio of temperature and humidity, used to reflect the combined influence of temperature and humidity on color change;

[0020] After calculating C pred (t), f(C i (t), E i (t)) needs to be generated, and the time series data of food color change needs to be analyzed in combination with food ingredient analysis technology, considering the influence of environmental conditions such as temperature and humidity, correcting the predicted value of color change to ensure the accuracy of the prediction;

[0021] f(C i (t), E i (t)) = C i (t) · (1 + g · sin(2π · F(E i (t)) + h · cos(2π · G(E i (t)))

[0022] where C i (t) represents the normalized color data of the i-th sample at time t; E i (t) represents the environmental condition data of the i-th sample at time t; g is a coefficient related to the periodic change of environmental conditions, used to adjust the influence of periodic change of environmental conditions on color change; h is a coefficient related to another periodic change of environmental conditions, used to adjust the influence of different periodic changes on color change; F(E i (t)) represents a frequency function of environmental condition E i (t), used to capture the periodic influence of environmental condition change on food color change; G(E i (t)) represents another frequency function of environmental condition E i (t), used to capture the different periodic influence of environmental condition change on food color change;

[0023] After calculating f(C i (t), E i (t)), a food color change prediction model needs to be generated, and the main trend, periodic characteristics of food color change and the influence of environmental conditions need to be integrated to optimize the model parameters using time series analysis method and food ingredient analysis technology, and finally generate an accurate model to predict food color change.

[0024] Optionally, the time series data of food color change is processed for trend and periodicity feature analysis based on the food color change prediction model, the color stability of food under different environmental conditions is evaluated, and a color change feature evaluation result is obtained, including:

[0025] The time series data of food color change is processed for trend analysis based on the food color change prediction model, the main trend direction and change rate of color change are identified, and color change trend information is obtained;

[0026] The time series data of food color change is processed for periodicity feature analysis using the color change trend information and Fourier transform technology, the periodicity mode of color change is extracted, and color change periodicity feature information is obtained;

[0027] According to the color change trend information and color change periodicity feature information, the color stability of food under different environmental conditions is evaluated by comparing the color change of food under different environmental conditions, and a color stability evaluation index is generated;

[0028] The color change characteristics of food under different environmental conditions are comprehensively evaluated using the color stability evaluation index and the correlation analysis of food ingredients and environmental conditions, and a color change feature evaluation result is obtained.

[0029] Optionally, the color change trend analysis report is generated by integrating the main trend, periodicity feature, and color change rate and amplitude information of food color change using the color change feature evaluation result, data visualization technology, and statistical analysis method, including:

[0030] The color change trend graph is obtained by generating a visual chart of food color change using the color change feature evaluation result and data visualization technology, which intuitively displays the main trend and periodicity feature of food color change;

[0031] Based on the color change trend graph, the rate and amplitude of food color change are quantitatively analyzed using statistical analysis methods, the severity and change law of food color change are evaluated, and color change rate and amplitude analysis results are obtained;

[0032] According to the color change rate and amplitude analysis results, the color stability of food under different environmental conditions is comprehensively evaluated, and the key factors affecting food color change are extracted, and a color stability comprehensive evaluation report is obtained;

[0033] The color change trend analysis report is generated by integrating the main trend, periodicity feature, and color change rate and amplitude information of food color change using the color stability comprehensive evaluation report.

[0034] Optionally, according to the color change trend analysis report, the decay rate and possible deterioration point of food freshness are predicted by using a long short-term memory network based on time series analysis combined with food ingredient analysis technology and a gray prediction model, and a food shelf life evaluation result is generated, including:

[0035] According to the color change trend analysis report, the main trend, periodic characteristics, change rate and amplitude information of food color change are extracted, and a time series data set of food color change is constructed;

[0036] Using the time series data set of food color change, combined with food ingredient analysis technology, a long short-term memory network is used to model and process the decay rate of food freshness, and a food freshness decay rate prediction model is generated;

[0037] Based on the food freshness decay rate prediction model, a gray prediction model is applied to supplement the prediction of the time series data of food color change, to enhance the accuracy and robustness of the prediction, and to obtain a comprehensive prediction result of food freshness;

[0038] Using the comprehensive prediction result of food freshness, the key nodes that may indicate deterioration in the food color change process are identified, the deterioration risk of food under different conditions is evaluated, and a food deterioration point prediction report is generated;

[0039] According to the food deterioration point prediction report, combined with food ingredient characteristics and historical data, the shelf life of food is comprehensively evaluated, and a food shelf life evaluation result is generated.

[0040] Optionally, based on the food freshness decay rate prediction model, a gray prediction model is applied to supplement the prediction of the time series data of food color change, to enhance the accuracy and robustness of the prediction, and to obtain a comprehensive prediction result of food freshness, including:

[0041] Before calculating F gray (t), the time series data of food color change needs to be preprocessed to extract environmental condition data, combined with food freshness decay rate prediction model C pred (t) and food ingredient analysis technology, to provide input for the supplementary prediction of the gray prediction model;

[0042]

[0043] Wherein, F gray(t) represents the food freshness supplementary prediction value obtained by applying the grey prediction model at time t, a is the weight coefficient of the grey prediction model, GM(1, 1)(t) represents the prediction value of the grey prediction model GM(1, 1) based on time t, β is an adjustment coefficient related to the food freshness prediction value, γ is a sensitivity coefficient related to the food color change prediction value, C pred (t) represents the food color change value predicted at time t, δ is a reference value of the food color change for adjusting the sensitivity of the prediction, λ is a coefficient related to the periodic change of the ambient temperature, ω is the angular frequency of the ambient temperature change, T(t) represents the ambient temperature at time t, and φ is the phase angle of the ambient temperature change.

[0044] After calculating F gray (t), C comp (t) is generated by fusing F gray (t) and the food color change prediction value C pred (t), adjusting the weight coefficients of both, and introducing the influence of the ambient humidity and the food packaging characteristics to generate a more accurate food freshness prediction value.

[0045]

[0046] C comp (t) represents the food freshness value obtained by comprehensive prediction at time t, θ is the weight coefficient of the food color change prediction value, C pred (t) represents the food color change value predicted at time t, F gray (t) represents the food freshness supplementary prediction value obtained by applying the grey prediction model at time t, μ is an adjustment coefficient related to the ambient humidity and the food packaging characteristics, H(t) represents the ambient humidity at time t, and P(t) represents the food packaging characteristic parameter at time t, which is used to reflect the influence of the packaging on the food freshness.

[0047] After calculating C comp (t), the comprehensive prediction result of the food freshness is obtained by combining the food ingredient characteristics and the historical data, evaluating the decay rate and possible deterioration points of the food freshness, identifying the key deterioration nodes, evaluating the deterioration risk, and finally generating the comprehensive prediction result of the food freshness.

[0048] Optionally, the food freshness decay rate prediction model is generated by modeling and processing the food freshness decay rate through a long short-term memory network by combining a food ingredient analysis technique, including:

[0049] Based on the time series data set of the color change of the food, combined with the food ingredient information provided by the food ingredient analysis technology, the data is preprocessed, including missing value filling, outlier processing and data standardization, to ensure that the data quality meets the modeling requirements;

[0050] Using the preprocessed food color change time series data, a long short-term memory network model is constructed, which can capture the long-term dependence relationship in the time series and effectively reflect the change rule of food freshness over time;

[0051] In the long short-term memory network model, the food ingredient information is used as auxiliary input, which acts on the model training process together with the time series data, enhances the understanding ability of the model to the change characteristics of food freshness, and improves the prediction accuracy;

[0052] By continuously iterating the long short-term memory network model, the model performance is evaluated using cross-validation technology to ensure the generalization ability of the model under different time periods and different food ingredient conditions, and finally a food freshness decay rate prediction model is generated, which can accurately predict the food freshness decay rate.

[0053] Optionally, using the food shelf life evaluation result, an intelligent logistics scheduling system driven by a multi-objective optimization algorithm is used to automatically plan and process the whole process path of food from production to consumption, dynamically adjust the storage conditions and transportation time window, and generate a logistics scheme, including:

[0054] Using the food shelf life evaluation result, a multi-objective optimization problem of food logistics path planning is constructed to obtain an optimization objective set;

[0055] Based on the optimization objective set, a multi-objective optimization algorithm is used to automatically plan and process the whole process path of food from production to consumption to obtain a preliminary logistics path planning scheme;

[0056] According to the preliminary logistics path planning scheme, combined with the food shelf life evaluation result, the storage conditions of each storage point are dynamically adjusted, and the storage conditions at least include temperature and humidity, and an optimized storage condition setting is generated;

[0057] Using the optimized storage condition setting, the transportation time window is further optimized to generate a final logistics scheme.

[0058] In a second aspect, the embodiments of the present application provide a real-time monitoring and analysis system for food color change, including:

[0059] A construction module is configured to capture color change images of the food in real time by using high-sensitivity imaging equipment from multiple angles and different spectral ranges, collect temperature and humidity, and light intensity parameters of an environment in which the food is located by using an environment sensing device, and construct a multi-modal food color dataset;

[0060] A correction module is configured to perform dynamic correction processing on food color changes based on the multi-modal food color dataset by combining a random forest algorithm under an ensemble learning framework and an adaptive color space mapping technology, and obtain a color change trend analysis report.

[0061] A prediction module is configured to perform prediction processing on a decay speed of food freshness and a possible deterioration point by using a long short-term memory network based on time series analysis in combination with food ingredient analysis technology and a gray prediction model according to the color change trend analysis report, and generate a food shelf life evaluation result.

[0062] An adjustment module is configured to perform automatic planning processing on a whole-process path of the food from production to consumption by using the food shelf life evaluation result, dynamically adjusting storage conditions and transportation time windows by using an intelligent logistics scheduling system driven by a multi-objective optimization algorithm, and generating a logistics scheme.

[0063] In the embodiments of the application, color change images of the food are captured in real time by using high-sensitivity imaging equipment from multiple angles and different spectral ranges, temperature and humidity, and light intensity parameters of an environment in which the food is located are collected by using an environment sensing device, and a multi-modal food color dataset is constructed. Dynamic correction processing is performed on food color changes based on the multi-modal food color dataset by combining a random forest algorithm under an ensemble learning framework and an adaptive color space mapping technology, and a color change trend analysis report is obtained. Prediction processing is performed on a decay speed of food freshness and a possible deterioration point by using a long short-term memory network based on time series analysis in combination with food ingredient analysis technology and a gray prediction model according to the color change trend analysis report, and a food shelf life evaluation result is generated. Automatic planning processing is performed on a whole-process path of the food from production to consumption by using the food shelf life evaluation result, dynamically adjusting storage conditions and transportation time windows by using an intelligent logistics scheduling system driven by a multi-objective optimization algorithm, and a logistics scheme is generated.

[0064] The technical scheme of the application has the following beneficial effects:

[0065] The present application captures the color change images of food in real time by using high-sensitivity imaging devices from multiple angles and different spectral ranges, and combines with the temperature, humidity, and light intensity parameters collected by the environmental perception device to construct a comprehensive multi-modal food color dataset. This method not only can more accurately capture the details of food color change, but also can effectively eliminate the color deviation caused by changes in light source, thereby providing more reliable color change trend analysis report. Based on this report, the method of combining random forest algorithm under ensemble learning framework and adaptive color space mapping technology is used to dynamically correct the food color change, further improving the accuracy of color change prediction. In addition, by combining long short-term memory network (LSTM) of time series analysis, food ingredient analysis technology and gray prediction model, the present application can accurately predict the decay rate of food freshness and its possible deterioration point, and generate scientific food shelf life evaluation results. This comprehensive prediction method overcomes the limitations of traditional methods that rely only on visual inspection and basic environmental parameter monitoring, greatly improving the accuracy and reliability of food freshness prediction. Finally, using the food shelf life evaluation results, the present application realizes the automatic planning of the whole process path of food from production to consumption through the intelligent logistics scheduling system driven by multi-objective optimization algorithm. The system can dynamically adjust the storage conditions and transportation time window to ensure that the food is stored and transported under the best conditions, thereby effectively extending the shelf life of the food, reducing waste, and improving the overall efficiency of the supply chain. This method solves the problem of lack of dynamic adjustment capability and inability to effectively extend the shelf life of food in existing solutions, and provides a more intelligent and efficient solution for the food industry.

[0066] Further, based on the multi-modal food color dataset, the present method first processes the color information in different spectral ranges through adaptive color space mapping to eliminate the influence caused by changes in light source, and then uses random forest algorithm combined with the color change pattern of food under specific environmental conditions to establish a prediction model. Then, the trend and periodicity of the time series color change data are analyzed to evaluate the color stability, and finally the main trend, periodic characteristics and change rate information are integrated to generate a comprehensive color change trend analysis report. This method processes food color data in detail and uses advanced machine learning techniques to build a prediction model, effectively improving the accuracy and reliability of food color change monitoring. It not only can more accurately identify subtle color changes caused by environmental factors, but also helps to understand the rules of food color evolution over time, providing strong data support for food quality control. In addition, by generating a detailed color change trend analysis report, relevant personnel can more intuitively understand the state change of food, so as to take timely measures to prevent food deterioration, ensure food safety, and optimize inventory management and logistics scheduling process.

[0067] Further, the method extracts the main trend, periodic characteristics, change rate and amplitude information of food color change according to the color change trend analysis report, and constructs a time series dataset. By using this dataset combined with food ingredient analysis technology, the long short-term memory network is used to model the decay rate of food freshness, and the gray prediction model is used for supplementary prediction to enhance the accuracy and robustness of the prediction, so as to generate the comprehensive prediction result of food freshness. Based on these prediction results, the key nodes of deterioration are identified, the deterioration risk under different conditions is evaluated, and finally the food shelf life evaluation result is generated. This method can more accurately predict the decay rate of food freshness and the possible deterioration points by combining advanced machine learning technology and food ingredient analysis, so as to provide more scientific and reliable food shelf life evaluation. This not only helps to discover and handle potential quality problems in time and reduce food waste, but also provides strong support for supply chain management to ensure the safety and quality of food in the whole circulation link. In addition, by identifying the key deterioration nodes, targeted measures can be taken to extend the shelf life of food, optimize inventory management and logistics scheduling, and improve overall operational efficiency.

[0068] These aspects or other aspects of the present application will be more apparent in the following description of the embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0069] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0070] Figure 1 A flow chart of a food color change real-time monitoring and analysis method provided by an embodiment of the present application;

[0071] Figure 2 A structural schematic diagram of a food color change real-time monitoring and analysis system provided by an embodiment of the present application;

[0072] Figure 3 A structural schematic diagram of a computing device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0073] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0074] In some of the flowcharts described in the specification and claims of the present application and in the above description of the drawings, a plurality of operations are included in a particular order, but it should be clearly understood that these operations can be performed in the order in which they appear in this text or in parallel, and the serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these flowcharts can include more or fewer operations, and these operations can be performed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this text are used to distinguish different messages, devices, modules, etc., and do not represent the order of precedence, nor do "first" and "second" represent different types.

[0075] The technical solutions in the embodiments of the present application will be described clearly and completely in the specification of the present application in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0076] Figure 1 A flowchart of a real-time monitoring and analysis method for food color change is provided in the embodiments of the present application, as shown in Figure 1 The method comprises:

[0077] 101, using high-sensitivity imaging devices from multiple angles and different spectral ranges to capture real-time color change images of the food, using environmental perception devices to collect temperature and humidity, light intensity parameters of the environment where the food is located, and constructing a multi-modal food color dataset;

[0078] The multi-modal food color dataset includes food color change images captured by high-sensitivity imaging devices from multiple angles and different spectral ranges (such as visible light, near-infrared, etc.), and temperature and humidity, light intensity and other parameters collected by environmental perception devices. These data are used to comprehensively monitor the color change of the food and the environmental conditions in which it is located.

[0079] In the embodiments of the present application, during the production, storage or transportation of food, high-sensitivity imaging devices are used to capture real-time color change images of the food from different angles and spectral ranges, and environmental perception devices are used to record temperature and humidity, light intensity and other information of the environment where the food is located. These images and environmental parameters together constitute a multi-modal dataset, providing basic data for subsequent color change analysis.

[0080] Suppose a food processing company needs to monitor the color changes of a batch of fresh fruits. First, multiple high-sensitivity cameras are installed at key positions on the fruit production line, each equipped with multiple spectral filters that can capture both visible light and near-infrared images. In addition, environmental sensors are deployed in the same area to monitor and record the temperature, humidity, and light intensity of the area in real time. Each time the fruit passes through the camera, the system automatically triggers the shooting and stores the image together with the corresponding environmental data into the database, forming a multi-modal food color dataset containing timestamps, images, and environmental parameters.

[0081] 102. Based on the multi-modal food color dataset, the dynamic correction processing of food color change is performed by combining the random forest algorithm under the integrated learning framework with the adaptive color space mapping technology, and a color change trend analysis report is obtained.

[0082] The dynamic correction processing is based on the multi-modal food color dataset, which uses the method of combining the random forest algorithm under the integrated learning framework with the adaptive color space mapping technology to eliminate the color deviation caused by light source changes, obtain standardized color data, and then model the prediction of food color change, and finally generate a color change trend analysis report.

[0083] In this step, by applying the adaptive color space mapping technology to the images in the multi-modal food color dataset, the color deviation caused by light source changes is eliminated, and then the random forest algorithm is used to combine the color change patterns of food under different environmental conditions to establish a prediction model. Based on this model, the time series data of food color change is analyzed for trend and periodicity characteristics, the stability of food color is evaluated, and finally a color change trend analysis report is integrated and generated.

[0084] Optionally, the dynamic correction processing of food color change based on the multi-modal food color dataset by combining the random forest algorithm under the integrated learning framework with the adaptive color space mapping technology in step 102 to obtain a color change trend analysis report includes:

[0085] The multi-modal food color dataset is used to perform adaptive color space mapping processing on color information in different spectral ranges, eliminate color deviation caused by light source changes, and obtain standardized color data; based on the standardized color data, a food color change prediction model is generated by combining a random forest algorithm with color change patterns of food under different environmental conditions; according to the food color change prediction model, trend and periodicity feature analysis processing is performed on time series data of food color change, color stability of food under different environmental conditions is evaluated, and color change characteristic evaluation results are obtained; using the color change characteristic evaluation results, main trends, periodic characteristics, and color change rate and amplitude information of food color change are integrated through data visualization technology and statistical analysis methods, and a color change trend analysis report is generated.

[0086] According to the food color change prediction model, trend and periodicity feature analysis processing is performed on time series data of food color change, color stability of food under different environmental conditions is evaluated, and color change characteristic evaluation results are obtained, including:

[0087] According to the food color change prediction model, trend and periodicity feature analysis processing is performed on time series data of food color change, color stability of food under different environmental conditions is evaluated, and color change characteristic evaluation results are obtained, including:

[0088] According to the food color change prediction model, trend and periodicity feature analysis processing is performed on time series data of food color change, color stability of food under different environmental conditions is evaluated, and color change characteristic evaluation results are obtained, including:

[0089] The color change trend analysis report is generated by using the color change trend graph and the color change rate and amplitude analysis result.

[0090] In the embodiments of the present application, first, the multi-modal food color data set is processed by adaptive color space mapping to eliminate color deviation caused by light source changes, and standardized color data is obtained; second, based on the standardized color data, a food color change prediction model is established by combining the color change patterns under different environmental conditions using the random forest algorithm; third, the time series data of food color change is analyzed using the model to identify the main trend direction and change rate, and the Fourier transform technique is used to extract the periodic characteristics to evaluate the color stability; finally, the main trend, periodic characteristics, and change rate and amplitude information are integrated by data visualization technology and statistical analysis methods to generate a color change trend analysis report.

[0091] Suppose a food processing enterprise needs to monitor the color change of a batch of fresh strawberries. First, capture the color change images of the strawberries from multiple angles and different spectral ranges, and record the environmental temperature, humidity, and light intensity parameters to construct a multi-modal food color data set; second, apply adaptive color space mapping technology to the images in the data set to eliminate color deviation caused by different light sources and obtain standardized color data; third, use the random forest algorithm to combine the color change patterns of the strawberries under different temperature, humidity, and light conditions to establish a food color change prediction model; fourth, based on the model, analyze the time series data of the color change of the strawberries to identify the main trend direction and change rate of the color change, and use the Fourier transform technique to extract the periodic patterns of the color change to evaluate the color stability of the strawberries under different environmental conditions; finally, generate a visual chart of the color change of the strawberries using data visualization technology to visually display the main trend and periodic characteristics, and use statistical analysis methods to quantitatively analyze the rate and amplitude of the color change to evaluate the change law, and finally generate a detailed color change trend analysis report to provide scientific basis for subsequent freshness prediction and logistics scheduling; through the above steps, the enterprise can more accurately monitor and manage the color change of the strawberries to ensure product quality and supply chain efficiency.

[0092] The present application considers that the construction of the food color change prediction model is to accurately predict the color change trend of the food under different environmental conditions, so as to evaluate the freshness and shelf life of the food. The model combines standardized color data, environmental condition data and food ingredient analysis technology, and is comprehensively modeled by random forest algorithm and adaptive color space mapping technology.

[0093] Optionally, the food color change prediction model is generated by combining the random forest algorithm with the color change mode of the food under different environmental conditions based on the standardized color data, and includes:

[0094] Before calculating C pred (t), the multi-modal food color data set needs to be preprocessed to eliminate color deviation caused by light source change by adaptive color space mapping technology, so as to obtain standardized color data and prepare for constructing a random forest model for predicting food color change;

[0095]

[0096] Wherein, C pred (t) represents the predicted food color change value at time t, w i is the weight of the i-th tree in the random forest, reflecting the contribution of each tree to the final prediction result; C i (t) represents the standardized color data of the i-th sample at time t; E i (t) represents the environmental condition data of the i-th sample at time t; N is the number of samples; α is an adjustment coefficient related to environmental temperature, used to adjust the influence of temperature on color change; β is an adjustment coefficient related to environmental humidity, used to adjust the influence of humidity on color change; T(t) represents the environmental temperature at time t; H(t) represents the environmental humidity at time t; δ is an adjustment coefficient related to the ratio of temperature and humidity, used to reflect the comprehensive influence of temperature and humidity on color change;

[0097] After calculating C pred (t), in order to generate f(C i (t), E i (t)), the time series data of food color change needs to be analyzed, combined with food ingredient analysis technology, considering the influence of environmental conditions such as temperature and humidity, correcting the predicted value of color change, to ensure the accuracy of the prediction;

[0098] f(C i (t), E i (t)) = C i (t)·(1+γ·sin(2π·F(E i(t)) + η · cos(2π · G(E i (t))

[0099] where C i (t) denotes the normalized color data of the i-th sample at time t; E i (t) denotes the environmental condition data of the i-th sample at time t; γ is a coefficient related to the periodic change of environmental conditions, used to adjust the influence of the periodic change of environmental conditions on color change; η is a coefficient related to another periodic change of environmental conditions, used to adjust the influence of different periodic changes on color change; F(E i (t)) represents the frequency function of environmental condition E i (t) at time t, used to capture the periodic influence of environmental condition change on food color change; G(E i (t)) represents another frequency function of environmental condition E i (t) at time t, used to capture the different periodic influence of environmental condition change on food color change;

[0100] After calculating f(C i (t), E i (t)), to generate a food color change prediction model, the main trend, periodic characteristics of food color change and the influence of environmental conditions need to be integrated, and time series analysis method and food ingredient analysis technology are used to optimize model parameters, finally an accurate food color change prediction model is generated.

[0101] This formula aims to accurately predict the color change of food by integrating learning methods (such as random forest) combined with environmental conditions (such as temperature, humidity and their periodic changes). C pred (t) improves the stability and accuracy of prediction by integrating the prediction results of multiple trees and adjusting the influence of temperature and humidity; f(C i (t), E i (t)) captures the complex patterns of color change by considering the periodic changes of normalized color data and environmental conditions. These designs ensure that the model can fully reflect the color change trend of food under different environmental conditions.

[0102] The following briefly introduces the reasons for the design of each term of the formula:

[0103]

[0104] Random forest contribution term This term represents the contribution of each tree in the random forest to the final prediction result, and improves the stability and accuracy of prediction through ensemble learning method;

[0105] Temperature influence term This term considers the effect of temperature on color change, where an exponential function is used to adjust for the non-linear effect of temperature change on color change;

[0106] Humidity to temperature ratio effect term This term considers the effect of the relative relationship between humidity and temperature on color change, a logarithmic function is used to capture the change in this ratio relationship;

[0107] The following briefly introduces the way to obtain each parameter of the formula:

[0108] w i The weight of each tree is obtained through the random forest training process; f(C i (t), E i (t)): calculated according to the formula f(C i (t), E i (t)) = C i (t)·(1+γ·sin(2π·F(E i (t)))+η·cos(2π·G(E i (t)))); α, β, δ are set by historical data fitting or expert experience; T(t) and H(t) are temperature and humidity data collected in real time by environmental sensing devices:

[0109] The following briefly introduces the reason for designing each term of the formula:

[0110] f(C i (t), E i (t)) = C i (t)·(1+γ·sin(2π·F(E i (t)))+η·cos(2π·G(E i (t)))

[0111] Basic color data term: C i (t): this term represents the normalized color data of the ith sample at time t, which is the basis for prediction;

[0112] Environmental periodicity effect term γ·sin(2π·F(E i (t)): this term considers the effect of periodic changes in environmental conditions on color change, and a sine function is used to capture the periodic changes;

[0113] Environmental periodicity effect term η·cos(2π·G(E i (t)): this term considers the effect of another periodic change on color change, and a cosine function is used to capture different periodic changes;

[0114] The following briefly introduces the way to obtain each parameter of the formula:

[0115] C i (t) normalized color data after adaptive color space mapping;E i (t) environmental condition data collected in real time by environmental perception device;γ,η fitted from historical data or set by expert experience;F(E i (t)) and G(E i (t)) frequency functions extracted from environmental condition data by Fourier transform or other methods;

[0116] Suppose a food processing enterprise needs to predict the color change of a batch of fresh fruits. First, the color change images of the fruits and the environmental temperature and humidity data are collected by high-sensitivity imaging equipment and environmental perception devices; second, the data is preprocessed to eliminate the deviation caused by light source changes to obtain normalized color data; then, a prediction model is constructed using the random forest algorithm combined with environmental condition data.

[0117] Example calculation

[0118] Assume N=5, the weight of each tree w1=0.2, w2=0.3, w3=0.1, w4=0.2, w5=0.2;

[0119] Normalized color data C i (t) are 0.8, 0.7, 0.9, 0.6, 0.8, respectively;

[0120] Environmental condition data E i (t) includes temperature T(t)=25℃ and humidity H(t)=60%;

[0121] Parameter settings:α=0.1,β=0.01,δ=0.2,γ=0.05,η=0.03;

[0122] Frequency functions F(E i (t))=0.01 and G(E i (t))=0.02;

[0123]

[0124] C pred (t)=0.71+0.00006+(-0.05)≈0.66006;

[0125] Through the above calculation, it is predicted that under the current environmental conditions, the color change value of the fruit is 0.66006. Assuming that the threshold for color change is set to 0.7, since the result 0.6606 is less than the set threshold, it indicates that under the given temperature and humidity conditions, the color change of the fruit is relatively slow, maintaining good freshness. Enterprises can optimize the storage and transportation conditions according to this prediction result to ensure that the fruit reaches the consumers in the best state.

[0126] By combining multi-modal food color datasets and environmental conditions, using random forest algorithms and adaptive color space mapping techniques, this method can more accurately predict the color change trend of food. This not only improves the accuracy of food freshness prediction, but also helps enterprises optimize environmental conditions during storage and transportation to ensure that food reaches consumers in the best state, thereby reducing the risk of food spoilage, extending shelf life, and improving overall supply chain efficiency.

[0127] 103、According to the color change trend analysis report, using a long short-term memory network based on time series analysis combined with food composition analysis technology and a gray prediction model, the decay rate of food freshness and possible spoilage points are predicted and processed to generate food shelf life evaluation results;

[0128] Based on the color change trend analysis report, using a long short-term memory network (LSTM) combined with food composition analysis technology and a gray prediction model, the decay rate of food freshness and possible spoilage points are predicted to generate food shelf life evaluation results.

[0129] In this step, the main trend, periodic characteristics, change rate and amplitude information in the color change trend analysis report are extracted to construct a time series dataset. Use LSTM to model the decay rate of food freshness, and then apply the gray prediction model for supplementary prediction to enhance the accuracy and robustness of the prediction. According to the comprehensive prediction results, identify the critical nodes of spoilage, evaluate the spoilage risk, and finally generate the food shelf life evaluation results.

[0130] Optionally, step 103, according to the color change trend analysis report, using a long short-term memory network based on time series analysis combined with food composition analysis technology and a gray prediction model, the decay rate of food freshness and possible spoilage points are predicted and processed to generate food shelf life evaluation results, including:

[0131] According to the color change trend analysis report, the main trend, periodic characteristics, change rate and amplitude information of the food color change are extracted, and a time series data set of food color change is constructed; by using the time series data set of food color change, combining food ingredient analysis technology, the decay rate of food freshness is modeled and processed by long short-term memory network, and a food freshness decay rate prediction model is generated; based on the food freshness decay rate prediction model, a gray prediction model is applied to supplement the prediction of the time series data of food color change, the accuracy and robustness of the prediction are enhanced, and a comprehensive prediction result of food freshness is obtained; by using the comprehensive prediction result of food freshness, the key nodes that may indicate deterioration in the food color change process are identified, the deterioration risk of food under different conditions is evaluated, and a food deterioration point prediction report is generated; according to the food deterioration point prediction report, the shelf life of food is comprehensively evaluated by combining food ingredient characteristics and historical data, and a food shelf life evaluation result is generated.

[0132] Among them, the use of the time series data set of food color change, combining food ingredient analysis technology, modeling and processing the decay rate of food freshness by long short-term memory network, generating food freshness decay rate prediction model, including:

[0133] Based on the time series data set of food color change, combining the food ingredient information provided by the food ingredient analysis technology, the data is preprocessed, including missing value filling, outlier processing and data standardization, to ensure that the data quality meets the modeling requirements; by using the preprocessed food color change time series data, a long short-term memory network model is constructed, which can capture the long-term dependence relationship in the time series and effectively reflect the change law of food freshness with time; in the long short-term memory network model, the food ingredient information is used as auxiliary input, which acts on the model training process together with the time series data, enhances the understanding ability of the model to the change characteristics of food freshness, and improves the prediction accuracy; by continuously iterating the long short-term memory network model, the model performance is evaluated using cross-validation technology to ensure the generalization ability of the model under different time periods and different food ingredient conditions, and finally a food freshness decay rate prediction model is generated, which can accurately predict the decay rate of food freshness.

[0134] In the embodiments of the present application, first, the main trend, periodic characteristics, change rate and amplitude information of the color change of the food are extracted from the color change trend analysis report to construct a time series dataset; second, the data is preprocessed, including missing value filling, outlier processing and data standardization, to ensure that the data quality meets the modeling requirements; third, a long short-term memory network model is constructed using the preprocessed data, which can capture the long-term dependence in the time series, effectively reflect the change law of the freshness of the food over time, and take the food ingredient information as auxiliary input to improve the prediction accuracy of the model; finally, the model is iteratively trained and the performance is evaluated using cross-validation techniques to ensure the generalization ability of the model under different time periods and different food ingredient conditions, and a food freshness decay rate prediction model is generated.

[0135] Suppose a food processing enterprise needs to monitor the freshness of a batch of fresh milk. First, the main trend, periodic characteristics, change rate and amplitude information of the color change of the milk are extracted from the color change trend analysis report to construct a time series dataset; second, the data is preprocessed, including missing value filling, outlier processing and data standardization, to ensure that the data quality meets the modeling requirements; third, a long short-term memory network model is constructed using the preprocessed data, which can capture the long-term dependence in the time series, effectively reflect the change law of the freshness of the milk over time, and take the milk ingredient information (such as fat content, protein content, etc.) as auxiliary input, which together with the time series data acts on the model training process to improve the prediction accuracy of the model; then, the model is iteratively trained and the model performance is evaluated using cross-validation techniques to ensure the generalization ability of the model under different time periods and different milk ingredient conditions, and finally a prediction model is generated that can accurately predict the decay rate of the freshness of the milk; then, based on the model, a grey prediction model is applied to supplement the prediction of the time series data of the color change of the milk, enhancing the accuracy and robustness of the prediction, and obtaining the comprehensive prediction result of the freshness of the milk; the comprehensive prediction result is used to identify the key nodes in the color change process of the milk that may indicate spoilage, evaluate the spoilage risk of the milk under different conditions, and generate a milk spoilage point prediction report; finally, the shelf life of the milk is evaluated comprehensively based on the milk ingredient characteristics and historical data to generate a shelf life evaluation result; through the above steps, the enterprise can more accurately predict the decay rate and spoilage point of the milk, thereby optimizing inventory management and logistics scheduling, ensuring product quality and consumer safety.

[0136] The present application considers that in food freshness prediction, combining the grey prediction model and the time series data of food color change can enhance the accuracy and robustness of the prediction. By introducing environmental condition data, food ingredient analysis technology and packaging characteristic parameters, the freshness change trend of the food can be more comprehensively evaluated, thereby generating a comprehensive prediction result.

[0137] Optionally, based on the food freshness decay rate prediction model, a grey prediction model is applied to supplementally predict the time series data of food color change, to enhance the accuracy and robustness of the prediction, and obtain a comprehensive prediction result of food freshness, comprising:

[0138] Before calculating F gray (t), the time series data of food color change needs to be preprocessed, environmental condition data is extracted, and the food freshness decay rate prediction model C pred (t) and food ingredient analysis technology are combined to provide input for the supplementally prediction of the grey prediction model;

[0139]

[0140] Wherein, F gray (t) represents the supplementally predicted value of food freshness at time t by applying the grey prediction model, α is the weight coefficient of the grey prediction model, GM(1, 1)(t) represents the prediction value of the grey prediction model GM(1, 1) based on time t, β is an adjustment coefficient related to the food freshness prediction value, γ is a sensitivity coefficient related to the food color change prediction value, C pred (t) represents the predicted value of food color change at time t, δ is a reference value of food color change for adjusting the sensitivity of the prediction, λ is a coefficient related to the periodic change of environmental temperature, ω is the angular frequency of the change of environmental temperature, T(t) represents the environmental temperature at time t, and φ is the phase angle of the change of environmental temperature;

[0141] After calculating F gray (t), in order to generate C comp (t), F gray (t) and the food color change prediction value C pred (t) need to be fused, the weight coefficients of the two are adjusted, and the influences of environmental humidity and food packaging characteristics are introduced, to generate a more accurate food freshness prediction value;

[0142]

[0143] Wherein, C comp (t) represents the food freshness value obtained by comprehensive prediction at time t, θ is the weight coefficient of the food color change prediction value, C pred (t) represents the predicted value of food color change at time t, and Fgray (t) represents the supplementary prediction value of food freshness obtained by applying the grey prediction model at time t, μ is an adjustment coefficient related to environmental humidity and food packaging characteristics, H(t) represents the environmental humidity at time t, and P(t) represents the food packaging characteristic parameter at time t, reflecting the impact of packaging on food freshness;

[0144] After calculating C comp (t), the attenuation speed of food freshness and possible deterioration points need to be evaluated, key deterioration nodes need to be identified, and deterioration risks need to be assessed to generate a comprehensive prediction result of food freshness.

[0145] This formula aims to enhance the accuracy and robustness of food freshness prediction by combining the grey prediction model and time series data of food color change. Specifically, this formula aims to comprehensively consider the impact of environmental conditions, food ingredient analysis technology, and packaging characteristics on food freshness, thereby generating a more comprehensive and reliable comprehensive prediction result to help enterprises make more accurate decisions in supply chain management.

[0146] The following briefly introduces the reasons for the design of each term of the formula:

[0147]

[0148] Grey prediction model term α·GM(1, 1)(t): use the grey prediction model GM(1, 1) to make supplementary prediction on time series data to improve the accuracy of prediction;

[0149] Food freshness prediction adjustment term Adjust the food freshness prediction value through the Sigmoid function to make it more smooth and have nonlinear characteristics;

[0150] Environmental temperature periodicity influence term λ·sin(ω·T(t)+φ): consider the influence of periodic changes of environmental temperature on food freshness, use the sine function to capture this periodic change;

[0151] The following briefly introduces the way to obtain each parameter of the formula:

[0152] α is set by historical data fitting or expert experience; GM(1, 1)(t) is calculated by the grey prediction model GM(1, 1); β is set by historical data fitting or expert experience; γ is set by historical data fitting or expert experience; Cpre d (t) is calculated by the food color change prediction model; δ is set by historical data fitting or expert experience; λ is set by historical data fitting or expert experience; ωAngular frequency extracted from ambient temperature data by Fourier transform etc.; T(t) Ambient temperature data collected in real-time by environmental sensing devices; φ Phase angle extracted from ambient temperature data by Fourier transform etc.

[0153] The following briefly introduces the design reasons for each term of the formula:

[0154]

[0155] Food color change prediction term θ·C pred (t): Retains the main contribution of food color change prediction value, ensures the accuracy of the basic prediction;

[0156] Grey prediction model supplementary term (1-θ)·F gray (t): Fuses the results of the grey prediction model with the food color change prediction value, further improves the robustness of the prediction;

[0157] Environmental humidity and packaging property influence term Consider the influence of environmental humidity and food packaging properties on freshness, use the logarithmic function to capture the changes of these factors;

[0158] The following briefly introduces the way to obtain each parameter of the formula:

[0159] θ is set by historical data fitting or expert experience; Cpred(t) is calculated by the food color change prediction model; Fgra y (t) is calculated by the above formula; μ is set by historical data fitting or expert experience; H(t) is the environmental humidity data collected in real-time by environmental sensing devices; P(t) is obtained by food packaging property data;

[0160] Suppose a food processing enterprise needs to predict the freshness of a batch of fresh fruits. First, the color change images of the fruits and the environmental temperature and humidity data are collected by high-sensitivity imaging equipment and environmental sensing devices; second, the data is preprocessed to eliminate the deviation caused by light source changes to obtain standardized color data; then, a comprehensive prediction model is constructed using the food freshness decay rate prediction model and the grey prediction model.

[0161] Suppose α=0.5, β=0.3, γ=0.05, δ=0.7, λ=0.2, ω=0.1, φ=0.2;

[0162] Prediction value Cpred(t)=0.66006;

[0163] Ambient temperature T(t)=25℃;

[0164] The grey prediction model value GM(1,1)(t) = 0.7;

[0165] θ = 0.6, μ = 0.1;

[0166] Ambient humidity H(t) = 60%;

[0167] Packaging characteristic parameter P(t) = 0.8;

[0168] Substitute data for calculation

[0169]

[0170] Fgray(t)=0.35+0.3·0.475+0.2·0.999≈0.6425;

[0171]

[0172] C comp (t)=0.396036+0.257+0.1·2.041≈0.657146;

[0173] Based on the above calculations, the predicted freshness value of the fruit under the current environmental conditions is 0.657146. Assuming a freshness threshold of 0.7, the result of 0.657146 is less than this threshold, indicating that the fruit maintains good freshness under the given temperature, humidity, and packaging conditions. Businesses can use this prediction to optimize storage and transportation conditions to ensure that the fruit reaches consumers in optimal condition.

[0174] By combining a grey prediction model with time-series data on food color changes, this method can significantly improve the accuracy and robustness of food freshness prediction. This not only helps companies more accurately assess the rate of freshness decay and potential spoilage points of food, but also optimizes storage and transportation conditions, ensuring that food reaches consumers in optimal condition, thereby reducing the risk of food spoilage, extending shelf life, and improving overall supply chain efficiency.

[0175] 104. Using the food shelf-life assessment results, an intelligent logistics scheduling system driven by a multi-objective optimization algorithm is used to automatically plan the entire process path of food from production to consumption, dynamically adjust storage conditions and transportation time windows, and generate logistics solutions.

[0176] Based on the food shelf-life assessment results, an intelligent logistics scheduling system driven by a multi-objective optimization algorithm automatically plans the entire process path of food from production to consumption, dynamically adjusts storage conditions and transportation time windows, and ensures that food is distributed under optimal conditions to extend its shelf life.

[0177] In this step, the food shelf life evaluation results are used to build a multi-objective optimization problem of logistics path planning. A multi-objective optimization algorithm is used to automatically plan the whole process path from production to consumption. According to the planning results, the temperature and humidity conditions of each warehouse are dynamically adjusted, and the transportation time window is optimized to generate the optimal logistics scheme.

[0178] Optionally, in step 104, the food shelf life evaluation results are used to automatically plan the whole process path from production to consumption by an intelligent logistics scheduling system driven by a multi-objective optimization algorithm, dynamically adjust the storage conditions and transportation time window, and generate a logistics scheme, including:

[0179] The food shelf life evaluation results are used to build a multi-objective optimization problem of food logistics path planning to obtain an optimization target set. Based on the optimization target set, a multi-objective optimization algorithm is used to automatically plan the whole process path from production to consumption to obtain a preliminary logistics path planning scheme. According to the preliminary logistics path planning scheme, the storage conditions of each warehouse are dynamically adjusted based on the food shelf life evaluation results, and the storage conditions at least include temperature and humidity to generate optimized storage condition settings. The optimized storage condition settings are used to further optimize the transportation time window to generate the final logistics scheme.

[0180] In the embodiments of the present application, first, the food shelf life evaluation results are used to build a multi-objective optimization problem of food logistics path planning to obtain an optimization target set. Second, based on the optimization target set, a multi-objective optimization algorithm is used to automatically plan the whole process path from production to consumption to obtain a preliminary logistics path planning scheme. Third, according to the preliminary logistics path planning scheme, the storage conditions of each warehouse are dynamically adjusted based on the food shelf life evaluation results, and the storage conditions at least include temperature and humidity to generate optimized storage condition settings. Finally, the optimized storage condition settings are used to further optimize the transportation time window to generate the final logistics scheme.

[0181] Assuming that a food processing enterprise needs to transport a batch of fresh fruits from the production site to multiple retail points. First, based on the food shelf life evaluation results, a multi-objective optimization problem of food logistics path planning is constructed, and the optimization objectives are set as minimizing transportation time and cost, and maximizing food freshness. Second, based on these optimization objectives, a genetic algorithm is used to automatically plan the logistics path from the production site to each retail point, considering the transportation time, cost and possible deterioration risk of different paths, and a preliminary logistics path planning scheme is obtained. Third, based on the preliminary logistics path planning scheme, combined with the food shelf life evaluation results, the storage conditions such as temperature and humidity of each warehouse are dynamically adjusted, for example, a lower temperature and appropriate humidity are set in the warehouse with longer storage time to delay the deterioration speed of fruits, and the optimized storage condition setting is generated. Finally, the transportation time window is further optimized using the optimized storage condition setting to ensure that the fruits are transported under the best conditions, avoid high temperature transportation, reduce the risk of deterioration, and generate the final logistics scheme. Through the above steps, the enterprise can ensure that the fruits remain in the best state throughout the supply chain, optimize logistics cost and efficiency, and improve overall operational efficiency.

[0182] Figure 2 A structural diagram of a food color change real-time monitoring and analysis system is provided for the embodiments of the present application, as shown in Figure 2 The device comprises:

[0183] The construction module 21 is configured to capture color change images of the food in real time from high-sensitivity imaging devices at multiple angles and different spectral ranges, collect temperature and humidity, and light intensity parameters of the environment where the food is located using an environmental perception device, and construct a multi-modal food color dataset.

[0184] The correction module 22 is configured to dynamically correct the color change of the food based on the multi-modal food color dataset by combining the random forest algorithm under the integrated learning framework and the adaptive color space mapping technology, and obtain a color change trend analysis report.

[0185] The prediction module 23 is configured to predict the decay rate and possible deterioration point of the food freshness based on the color change trend analysis report by using the long short-term memory network based on time series analysis combined with the food ingredient analysis technology and the gray prediction model, and generate a food shelf life evaluation result.

[0186] The adjustment module 24 is configured to automatically plan the whole process path of the food from production to consumption by using the food shelf life evaluation result through the intelligent logistics scheduling system driven by the multi-objective optimization algorithm, dynamically adjust the storage conditions and transportation time window, and generate a logistics scheme.

[0187] Figure 2The food color change real-time monitoring and analysis system can perform Figure 1 The food color change real-time monitoring and analysis method of the embodiment has the same implementation principle and technical effects as the food color change real-time monitoring and analysis system. The specific operation modes of each module and unit of the food color change real-time monitoring and analysis system have been described in detail in the embodiment of the method, and will not be described here in detail.

[0188] In one possible design, Figure 2 The food color change real-time monitoring and analysis system of the embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device can include a storage component 31 and a processing component 32.

[0189] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32.

[0190] The processing component 32 is configured to capture color change images of the food in real time by using high-sensitivity imaging devices from multiple angles and different spectral ranges, collect temperature and humidity, and light intensity parameters of the environment where the food is located by using an environmental perception device, and construct a multi-modal food color dataset; based on the multi-modal food color dataset, dynamically correct the food color change by combining a random forest algorithm under an ensemble learning framework and an adaptive color space mapping technology, and obtain a color change trend analysis report; according to the color change trend analysis report, predict the decay rate of the food freshness and the possible deterioration point by using a long short-term memory network based on time series analysis combined with food ingredient analysis technology and a gray prediction model, and generate a food shelf life evaluation result; and use the food shelf life evaluation result to automatically plan the whole process path of the food from production to consumption by using an intelligent logistics scheduling system driven by a multi-objective optimization algorithm, dynamically adjust the storage conditions and the transportation time window, and generate a logistics scheme.

[0191] The processing component 32 can include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component can also be one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic elements, for executing the above method.

[0192] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component can be implemented by any type of volatile or nonvolatile storage devices or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic or optical disk.

[0193] Of course, the computing device can also necessarily include other components, such as an input / output interface, a display component, a communication component, etc.

[0194] The input / output interface provides an interface between the processing component and peripheral interface modules, which can be output devices, input devices, etc.

[0195] The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, etc.

[0196] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform, and the computing device can be a cloud server, and the processing component, the storage component, etc. can be a basic server resource rented or purchased from the cloud computing platform.

[0197] The embodiment of the application further provides a computer storage medium, which stores a computer program, and the computer program can implement the above-mentioned Figure 1 The embodiment shown provides a real-time monitoring and analysis method for food color change.

[0198] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-mentioned system, device and unit can refer to the corresponding process in the foregoing method embodiment, which will not be described here.

[0199] The device embodiment described above is only schematic, wherein the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment scheme. Those skilled in the art can understand and implement it without creative labor.

[0200] Those skilled in the art can clearly understand the implementation of the various embodiments by means of software and the necessary general hardware platform from the above description of the embodiments, and of course, the embodiments can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that contributes to the technical solutions can be embodied in the form of a software product. The computer software product can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the methods.

[0201] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacement for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for real-time monitoring and analysis of food color changes, characterized in that, include: By using high-sensitivity imaging equipment that captures color change images of food in real time from multiple angles and different spectral ranges, and by using environmental sensing devices to collect temperature, humidity and light intensity parameters of the environment in which the food is located, a multimodal food color dataset is constructed. Based on the multimodal food color dataset, the random forest algorithm under the ensemble learning framework is combined with the adaptive color space mapping technology to dynamically correct food color changes and obtain a color change trend analysis report. Based on the color change trend analysis report, the long short-term memory network based on time series analysis, combined with food component analysis technology and gray prediction model, is used to predict the rate of decay of food freshness and possible spoilage points, and generate food shelf life assessment results. Using the food shelf-life assessment results, an intelligent logistics scheduling system driven by a multi-objective optimization algorithm is used to automatically plan the entire process path of food from production to consumption, dynamically adjust storage conditions and transportation time windows, and generate logistics solutions. Specifically, based on the color change trend analysis report, the long short-term memory network based on time series analysis, combined with food component analysis technology and a grey prediction model, is used to predict the rate of freshness decay and possible spoilage points of the food, generating a food shelf-life assessment result, including: Based on the color change trend analysis report, extract the main trends, periodic characteristics, rate of change and magnitude information of food color changes, and construct a time series dataset of food color changes; Using the time series dataset of food color changes, combined with food composition analysis technology, a long short-term memory network is used to model the rate of food freshness decay, generating a food freshness decay rate prediction model. Based on the food freshness decay rate prediction model, a grey prediction model is applied to supplement the prediction of time series data of food color changes, thereby enhancing the accuracy and robustness of the prediction and obtaining a comprehensive prediction result of food freshness. Using the comprehensive prediction results of food freshness, key nodes that may indicate spoilage during the color change process of food are identified, the risk of food spoilage under different conditions is assessed, and a food spoilage point prediction report is generated. Based on the food spoilage point prediction report, combined with the food's component characteristics and historical data, a comprehensive assessment of the food's shelf life is conducted to generate a food shelf life assessment result.

2. The method according to claim 1, characterized in that, Based on the multimodal food color dataset, the random forest algorithm under the ensemble learning framework is combined with adaptive color space mapping technology to dynamically correct food color changes, resulting in a color change trend analysis report, including: Using the aforementioned multimodal food color dataset, adaptive color space mapping processing is performed on color information in different spectral ranges to eliminate color deviations caused by changes in light source and obtain standardized color data. Based on the standardized color data, a food color change prediction model is generated by combining the random forest algorithm with the color change patterns of food under different environmental conditions to predict and model food color changes. Based on the food color change prediction model, trend and periodic characteristics of the time series data of food color change are analyzed and processed to evaluate the color stability of food under different environmental conditions and obtain the color change characteristic evaluation results. Using the color change characteristic evaluation results, and through data visualization technology and statistical analysis methods, the main trends, periodic characteristics, and color change rate and magnitude information of food color changes are integrated to generate a color change trend analysis report.

3. The method according to claim 2, characterized in that, Based on the standardized color data, a food color change prediction model is generated by combining the color change patterns of food under different environmental conditions with a random forest algorithm, including: In calculating C pred Before (t), the multimodal food color dataset needs to be preprocessed. Adaptive color space mapping technology is used to eliminate color deviation caused by changes in light source to obtain standardized color data, which is used to prepare for building a random forest model to predict food color changes. Among them, C pred (t) represents the predicted food color change value at time t, w i C represents the weight of the i-th tree in the random forest, reflecting the contribution of each tree to the final prediction result; i (t) represents the standardized color data of the i-th sample at time t; E i (t) represents the environmental condition data of the i-th sample at time t; N is the number of samples; α is an adjustment coefficient related to ambient temperature, used to adjust the effect of temperature on color change; β is an adjustment coefficient related to ambient humidity, used to adjust the effect of humidity on color change; T(t) represents the ambient temperature at time t; H(t) represents the ambient humidity at time t; δ is an adjustment coefficient related to the temperature and humidity ratio, used to reflect the combined effect of temperature and humidity on color change; After calculating C pred After (t), to generate f(C) i (t),E i (t) requires analysis of time series data on food color changes, combined with food composition analysis techniques, taking into account the influence of environmental conditions such as temperature and humidity, to correct the predicted values ​​of color changes and ensure the accuracy of the prediction. f(C i (t),E i (t))=C i (t)·(1+γ·sin(2π·F(E i (t)))+η·cos(2π·G(E i (t)))) Among them, C i (t) represents the standardized color data of the i-th sample at time t; E i (t) represents the environmental condition data of the i-th sample at time t; γ is a coefficient related to the periodic changes in environmental conditions, used to adjust for the influence of periodic changes in environmental conditions on color change; η is a coefficient related to another periodic change in environmental conditions, used to adjust for the influence of different periodic changes on color change; F(E i (t) represents environmental condition E i The frequency function of G(t) is used to capture the periodic effect of changes in environmental conditions on food color changes; i (t) represents environmental condition E i Another frequency function of (t) is used to capture the different periodic effects of changes in environmental conditions on food color changes; After calculating f(C) i (t),E i After (t)), in order to generate a food color change prediction model, it is necessary to comprehensively consider the main trends, periodic characteristics and environmental conditions of food color change, optimize the model parameters using time series analysis methods and food component analysis techniques, and finally generate a model that accurately predicts food color changes.

4. The method according to claim 2, characterized in that, The process involves analyzing the trend and periodic characteristics of time-series data on food color changes based on the food color change prediction model, evaluating the color stability of food under different environmental conditions, and obtaining color change characteristic evaluation results, including: Based on the food color change prediction model, trend analysis is performed on the time series data of food color change to identify the main trend direction and rate of change of color change, and to obtain color change trend information. Using the color change trend information and Fourier transform technology, the time series data of food color change is analyzed for periodic features to extract the periodic pattern of color change and obtain the periodic feature information of color change. Based on the color change trend information and color change periodicity characteristic information, the color stability of food under different environmental conditions is evaluated by comparing the color changes of food under different environmental conditions, and a color stability evaluation index is generated. By using the aforementioned color stability assessment index, combined with the correlation analysis of food components and environmental conditions, the color change characteristics of food under different environmental conditions are comprehensively evaluated, and the color change characteristic assessment results are obtained.

5. The method according to claim 2, characterized in that, The evaluation results based on the color change characteristics are used to integrate the main trends, periodic characteristics, and color change rate and magnitude information of food color changes through data visualization technology and statistical analysis methods, generating a color change trend analysis report, including: Using the evaluation results of the color change characteristics, a visual chart of food color change is generated through data visualization technology to intuitively display the main trends and periodic characteristics of food color change, resulting in a color change trend chart. Based on the color change trend chart, combined with statistical analysis methods, the rate and magnitude of food color change are quantitatively analyzed to assess the intensity and pattern of food color change, and the analysis results of color change rate and magnitude are obtained. Based on the analysis results of the color change rate and amplitude, the color stability of food under different environmental conditions is comprehensively evaluated, the key factors affecting the color change of food are extracted, and a comprehensive color stability evaluation report is obtained. Using the aforementioned color stability comprehensive evaluation report, information on the main trends, periodic characteristics, and rate and magnitude of color changes in food is integrated to generate a color change trend analysis report.

6. The method according to claim 1, characterized in that, The aforementioned food freshness decay rate prediction model utilizes a grey prediction model to supplement the prediction of time-series data on food color changes, enhancing the accuracy and robustness of the prediction, and obtaining a comprehensive prediction result for food freshness, including: In calculating F gray Before (t), the time series data of food color change needs to be preprocessed, environmental condition data needs to be extracted, and the data needs to be combined with the food freshness decay rate prediction model C. pred (t) and food composition analysis techniques provide input for supplementary predictions in the grey prediction model; Among them, F gray (t) represents the supplementary predicted value of food freshness obtained by applying the grey prediction model at time t, α is the weight coefficient of the grey prediction model, GM(1,1)(t) represents the predicted value of the grey prediction model GM(1,1) based on time t, β is the adjustment coefficient related to the predicted value of food freshness, γ is the sensitivity coefficient related to the predicted value of food color change, and C pred (t) represents the predicted food color change value at time t, δ is the baseline value of food color change used to adjust the sensitivity of the prediction, λ is the coefficient related to the periodic change of ambient temperature, ω is the angular frequency of the ambient temperature change, T(t) represents the ambient temperature at time t, and φ is the phase angle of the ambient temperature change. After calculating F gray After (t), to generate C comp (t), F needs to be gray (t) and the predicted value of food color change C pred (t) is fused, the weight coefficients of the two are adjusted, and the influence of environmental humidity and food packaging characteristics is introduced to generate a more accurate food freshness prediction value; Among them, C comp (t) represents the food freshness value obtained from the comprehensive prediction at time t, θ is the weighting coefficient of the predicted food color change value, and C pred (t) represents the predicted food color change value at time t, F gray (t) represents the supplementary predicted value of food freshness obtained by applying the grey prediction model at time t, μ is the adjustment coefficient related to environmental humidity and food packaging characteristics, H(t) represents the environmental humidity at time t, and P(t) represents the food packaging characteristic parameter at time t, which is used to reflect the impact of packaging on food freshness. After calculating C comp After (t), in order to obtain a comprehensive prediction result of food freshness, it is necessary to combine the characteristics of food components and historical data to assess the rate of decline of food freshness and possible spoilage points, identify key spoilage nodes, assess spoilage risk, and finally generate a comprehensive prediction result of food freshness.

7. The method according to claim 1, characterized in that, The process of using the time-series dataset of food color changes, combined with food component analysis techniques, and modeling the rate of food freshness decay using a long short-term memory network to generate a food freshness decay rate prediction model includes: Based on the time series dataset of food color changes, and combined with the food composition information provided by food composition analysis technology, the data is preprocessed, including missing value imputation, outlier handling, and data standardization, to ensure that the data quality meets the modeling requirements. Using preprocessed time series data of food color changes, a long short-term memory network model was constructed. This long short-term memory network model can capture long-term dependencies in the time series and effectively reflect the changing pattern of food freshness over time. In the Long Short-Term Memory Network model, food composition information is used as an auxiliary input, working together with time series data in the model training process to enhance the model's understanding of the characteristics of food freshness changes and improve prediction accuracy. By iteratively training the Long Short-Term Memory network model and using cross-validation to evaluate its performance, the model's generalization ability under different time periods and food composition conditions is ensured, ultimately generating a food freshness decay rate prediction model that can accurately predict the rate of food freshness decay.

8. The method according to claim 1, characterized in that, The aforementioned intelligent logistics scheduling system, driven by a multi-objective optimization algorithm and utilizing the food shelf-life assessment results, automatically plans the entire process path of food from production to consumption, dynamically adjusts storage conditions and transportation time windows, and generates logistics solutions, including: Using the food shelf-life assessment results, a multi-objective optimization problem for food logistics route planning is constructed, resulting in a set of optimization objectives. Based on the aforementioned set of optimization objectives, a multi-objective optimization algorithm is used to automatically plan the entire process path of food from production to consumption, resulting in a preliminary logistics path planning scheme. Based on the preliminary logistics route planning scheme and the food shelf life assessment results, the storage conditions of each storage point are dynamically adjusted. The storage conditions include at least temperature and humidity, and optimized storage condition settings are generated. By utilizing the optimized warehousing conditions, the transportation time window is further optimized to generate the final logistics solution.

9. A real-time monitoring and analysis system for food color changes, used to execute the method for real-time monitoring and analysis of food color changes as described in any one of claims 1 to 8, characterized in that, include: The module is designed to capture images of food color changes in real time using high-sensitivity imaging devices from multiple angles and different spectral ranges, and to collect temperature, humidity, and light intensity parameters of the food's environment using environmental sensing devices, thereby constructing a multimodal food color dataset. The correction module is used to dynamically correct food color changes based on the multimodal food color dataset by combining the random forest algorithm under the ensemble learning framework with the adaptive color space mapping technology, and to obtain a color change trend analysis report. The prediction module is used to predict the rate of decay of food freshness and possible spoilage points based on the color change trend analysis report, using a long short-term memory network based on time series analysis combined with food component analysis technology and a gray prediction model, and generate food shelf life assessment results. The adjustment module is used to utilize the food shelf-life assessment results and, through a multi-objective optimization algorithm-driven intelligent logistics scheduling system, automatically plan the entire process path of food from production to consumption, dynamically adjust storage conditions and transportation time windows, and generate logistics solutions.

Images