In-transit monitoring management method and system for logistics transportation

Abstract

The invention belongs to the technical field of data processing, and particularly relates to an in-transit monitoring management method and system for logistics transportation, and the method comprises the steps: collecting transportation environment data and an agricultural product gray scale image in real time, and extracting agricultural product state indexes including the area, the gray scale and the texture change rate after preprocessing; deducing the state evolution trend of the agricultural products in combination with the predicted temperature and humidity at each moment in the whole transportation process; and based on a particle swarm optimization algorithm, constructing a state risk value fused by a degradation rate of an agricultural product state index and a temperature and humidity deviation degree, dynamically adjusting a particle inertia weight and a social learning factor, carrying out iterative optimization by taking a fitness function as a purpose, and outputting a transportation scheme including an optimal driving speed, temperature, humidity and path. According to the invention, the dynamic association between the environment and the quality of the agricultural products is realized, the intelligence and accuracy of cold chain regulation are improved, and the fresh-keeping effect and energy consumption of the agricultural products are balanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing. More particularly, the present application relates to an in-transit monitoring and management method and system for logistics transportation. BACKGROUND

[0002] Agricultural product cold chain logistics is a key link to ensure the quality of fresh agricultural products. The core goal is to delay the quality deterioration of agricultural products and extend the shelf life by environmental regulation (such as temperature and humidity control) during transportation.

[0003] However, the prior art still has many deficiencies in the in-transit monitoring and management of agricultural product transportation. Traditional cold chain transportation mostly only focuses on real-time collection and alarm of environmental parameters such as temperature and humidity, and lacks dynamic perception of the state of agricultural products. For example, some systems monitor the temperature and humidity of the carriage through sensors and trigger abnormal alarms, but cannot directly associate with the freshness change of agricultural products, and it is difficult to quantify the actual impact of environmental fluctuations on quality, resulting in that the control measures lag behind the quality deterioration process.

[0004] At the same time, the existing cold chain logistics scheme optimization is mostly based on fixed parameters or historical experience, such as preset refrigeration temperature, fixed distribution route, etc., without considering the dynamic coupling relationship between the environment and the state of agricultural products during transportation. For example, the traditional particle swarm optimization algorithm uses a global unified inertia weight, which is difficult to adjust the search strategy according to the real-time quality risk. In high-risk sections with high temperature and humidity, the best temperature control parameters are easily missed due to insufficient search precision; in low-risk sections with stable low temperature, the convergence is too slow, resulting in redundant energy consumption calculation, and finally the dynamic balance between preservation effect and energy cost cannot be achieved.

[0005] Therefore, there is an urgent need for an in-transit monitoring and management method for logistics transportation, which dynamically associates the changes of transportation environment and agricultural product state, and improves the intelligent and fine level of cold chain logistics. SUMMARY

[0006] To solve the above technical problems of the existing agricultural product cold chain logistics in-transit monitoring and management, the environment monitoring and quality perception are disconnected, and the dynamic coupling relationship between the environment and the state of agricultural products is not considered, resulting in lagging control and poor optimization effect, the present application provides a solution in the following aspects.

[0007] In a first aspect, the present application provides an in-transit monitoring and management method for logistics transportation, comprising: Real-time acquisition of transportation environment data and gray image of agricultural products, extraction of agricultural product state indicators according to the gray image; based on the particle swarm optimization algorithm, combining the predicted environmental data and agricultural product state indicators at each time of the whole transportation process, the state risk value of the transportation scheme corresponding to each particle is determined; the inertia weight and social learning factor of each particle are dynamically adjusted according to the state risk value, and the optimal cold chain transportation scheme including the optimal driving speed, set temperature and humidity and distribution path is output by iterative optimization taking the fitness function as the target.

[0008] The present application realizes the dynamic correlation of transportation environment and agricultural product state by real-time acquisition of environmental data and agricultural product image and extraction of agricultural product state indicators, determination of state risk value combining predicted whole-process environmental data and state indicators, dynamic adjustment of inertia weight and social learning factor of particles in particle swarm optimization algorithm, iterative optimization taking fitness function as the target and output of optimal transportation scheme, so that the optimization strategy can be flexibly adjusted according to the actual risk level, the perception accuracy of the quality change of agricultural products is improved, and the adaptability of the transportation scheme in different risk scenarios is enhanced, so that the freshness of agricultural products can be guaranteed while the energy consumption and path planning are optimized, the preservation effect and transportation cost are effectively balanced, and the intelligent and fine management level of cold chain logistics is significantly improved.

[0009] Preferably, the extraction of agricultural product state indicators according to the gray image comprises: instance segmentation of the gray image, extraction of the agricultural product region, realization of the agricultural product region matching between the current time and the initial time through SIFT key point matching and Hungarian algorithm, calculation of the area change rate, gray change rate and texture change rate of the agricultural product region between different frames of gray images, and fusion to obtain the agricultural product state characteristic indicators.

[0010] The present application realizes the dynamic correlation of transportation environment and agricultural product state by real-time acquisition of environmental data and agricultural product image and extraction of agricultural product state indicators, determination of state risk value combining predicted whole-process environmental data and state indicators, dynamic adjustment of inertia weight and social learning factor of particles in particle swarm optimization algorithm, iterative optimization taking fitness function as the target and output of optimal transportation scheme, so that the optimization strategy can be flexibly adjusted according to the actual risk level, the perception accuracy of the quality change of agricultural products is improved, and the adaptability of the transportation scheme in different risk scenarios is enhanced, so that the freshness of agricultural products can be guaranteed while the energy consumption and path planning are optimized, the preservation effect and transportation cost are effectively balanced, and the intelligent and fine management level of cold chain logistics is significantly improved.

[0011] Preferably, the area change rate is the average of the area shrinkage ratio of the agricultural product region corresponding to the current time and the initial time; the gray change rate is the absolute value average of the relative change of the average gray of the agricultural product region corresponding to the current time and the initial time; and the texture change rate is the absolute value average of the relative change of the entropy value of the gray co-occurrence matrix of the agricultural product region corresponding to the current time and the initial time.

[0012] Preferably, the state risk value obtaining method is: according to the difference between the predicted state characteristic index of the agricultural products at the transportation end under the transportation scheme represented by the particle and the state characteristic index at the current time, dividing by the remaining time from the current time to the transportation end, to obtain the deterioration rate of the state characteristic index of the agricultural products; and according to the deterioration rate, the temperature deviation degree and the humidity deviation degree of the whole transportation process of the transportation scheme, the state risk value is obtained by fusion.

[0013] The application obtains the deterioration rate by calculating the ratio of the difference between the predicted state characteristic index of the agricultural products at the transportation end and the state characteristic index at the current time and the remaining time, and obtains the state risk value by combining the temperature deviation degree and the humidity deviation degree of the whole transportation process, which can comprehensively reflect the synergistic effect of the quality decline speed of the agricultural products and the environmental control deviation, capture the dynamic trend of quality deterioration and take into account the cumulative effect of environmental factors, so that the risk assessment is more in line with the coupling relationship between the environment and the quality in the transportation process, provides accurate risk quantization basis for subsequent dynamic adjustment and optimization of parameters, avoids one-sidedness of single index evaluation, and improves the pertinence and reliability of the transportation scheme optimization.

[0014] Preferably, the temperature deviation degree is the normalized result of the average of the deviation proportion of the predicted temperature at each time compared with the optimal storage temperature interval of the agricultural products under the transportation scheme represented by the particle; and the humidity deviation degree is the normalized result of the average of the deviation proportion of the predicted humidity at each time compared with the optimal storage humidity interval of the agricultural products under the transportation scheme represented by the particle.

[0015] Preferably, the state risk value satisfies the expression: ; wherein, represents the state risk value of the i th particle at the j th iteration; represents the state risk value of the i th particle at the j th iteration; represents the deterioration rate of the state characteristic index of the agricultural products in the sliding time window under the transportation scheme represented by the i th particle; represents the deterioration rate of the state characteristic index of the agricultural products in the sliding time window under the transportation scheme represented by the i th particle; and represent the temperature deviation degree and the humidity deviation degree of the transportation scheme represented by the i th particle, respectively; and represent the temperature deviation degree and the humidity deviation degree of the transportation scheme represented by the i th particle, respectively; is the temperature deviation degree and the humidity deviation degree of the transportation scheme represented by the i th particle, respectively; is the temperature deviation degree and the humidity deviation degree of the transportation scheme represented by the i th particle, respectively;

[0016] ​The application can not only differentially weigh the influence of environmental factors and quality deterioration according to the sensitive characteristics of agricultural products to temperature and humidity, but also highlight the role of the dominant environmental stressor, so that the risk assessment is more in line with the preservation characteristics of different types of agricultural products, avoiding the problem of insufficient adaptability of uniform evaluation standards to different agricultural products, providing more targeted risk quantification basis for subsequent dynamic adjustment and optimization of parameters, and improving the accuracy of risk assessment and the adaptability of scheme optimization.

[0017] Preferably, the inertia weight satisfies the expression: ; wherein, represents the inertia weight of the i th particle at the j th iteration; and are the lower limit and the upper limit of the preset inertia weight, respectively; represents the state risk value of the i th particle at the j th iteration; is the decay coefficient.

[0018] The application dynamically adjusts the inertia weight through the state risk value. When the risk value is low, the inertia weight is close to the upper limit, and the particle maintains a large inertia to perform global exploration to quickly optimize. When the risk value is high, the inertia weight tends to the lower limit, and the particle enters a fine search mode to improve the local optimization accuracy. Meanwhile, the decay coefficient can adjust the sensitivity of the weight to the risk change, so that the particle swarm optimization can adaptively switch the search strategy according to the transportation risk level, avoiding the problem of fixed weight that the search is rough at high risk and converges slowly at low risk, and improving the adaptability and optimization efficiency of the algorithm to complex transportation scenarios.

[0019] Preferably, the social learning factor satisfies the expression: ; wherein, represents the social learning factor of the i th particle at the j th iteration; is the initial social learning factor; represents the state risk value of the i th particle at the j th iteration; is the risk influence decay coefficient.

[0020] ​​​​​​​​The application dynamically adjusts the social learning factor through the state risk value. When the risk value is low, the social learning factor is close to the initial value, and the particle is more inclined to follow the global optimal solution to speed up the convergence. When the risk value is high, the social learning factor decreases with the increase of the risk, and the particle reduces the dependence on the global optimal solution to enhance the autonomous exploration ability, avoiding the problem of local convergence caused by the blind following of the global solution by the particle in the high risk, while ensuring the convergence efficiency in the low risk, so that the particle swarm optimization can maintain search diversity and accurately focus on high-quality solutions in a complex dynamic transportation environment, and the robustness and effectiveness of the optimization are improved.

[0021] Preferably, the fitness function satisfies the expression: ; wherein, represents the comprehensive fitness value of the i th particle; represents the total refrigeration and dehumidification energy consumption of the transportation scheme corresponding to the i th particle; represents the predicted state characteristic index of the agricultural products at the arrival of the transportation terminal under the transportation scheme corresponding to the i th particle; and are the temperature deviation and humidity deviation of the transportation scheme represented by the i th particle, respectively; are weight coefficients.

[0022] In the second aspect, the application provides an in-transit monitoring and management system for logistics transportation, comprising a processor and a memory, and the memory stores computer program instructions which, when executed by the processor, implement the above-mentioned in-transit monitoring and management method for logistics transportation.

[0023] By adopting the above technical solution, the above-mentioned in-transit monitoring and management method for logistics transportation is generated into a computer program and stored in the memory to be loaded and executed by the processor, so as to manufacture a terminal device according to the memory and the processor, and facilitate use.

[0024] The application has the following advantages: (1) The application realizes the all-round upgrade of the in-transit management of the agricultural product cold chain logistics by constructing a closed-loop management mechanism of "real-time monitoring-dynamic perception-intelligent optimization".

[0025] (2) The application accurately extracts the state index reflecting the freshness of the agricultural products by collecting transportation environment data and agricultural product images in multiple dimensions, combining instance segmentation, cross-frame matching and multi-feature fusion technology, and solves the problems of disconnection between environment and quality and one-sided evaluation in traditional monitoring, and provides an objective quantitative basis for quality control. ​​​​​​

[0026] (3) The application introduces the state risk value concept, dynamically correlates the deterioration rate of agricultural products with the humidity deviation degree, and adaptively adjusts the inertia weight and social learning factor of the particle swarm optimization algorithm based on this, so that the optimization process can flexibly switch the search strategy according to the transportation risk level, focus on fine regulation in high-risk situations, and balance efficiency and energy consumption in low-risk situations, overcoming the poor adaptability of traditional static optimization parameters.

[0027] (4) The application balances the preservation effect, energy cost and environmental control precision, and outputs an optimal transportation scheme that can maximize the quality of agricultural products and realize reasonable resource allocation, significantly improving the intelligent and fine level of cold chain logistics, and providing an efficient solution for in-transit quality assurance and cost optimization of fresh agricultural products. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 is a flowchart schematically showing a method for in-transit monitoring and management of logistics transportation according to the application; Figure 2 is a flowchart schematically showing step S3 of the method for in-transit monitoring and management of logistics transportation according to the application. DETAILED DESCRIPTION

[0029] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0030] The specific embodiments of the application will be described in detail below with reference to the drawings.

[0031] The embodiments of the application disclose a method for in-transit monitoring and management of logistics transportation, referring to Figure 1 , comprising steps S1-S3: S1, real-time collection of environmental data and gray images of agricultural products during transportation.

[0032] It should be noted that during the transportation of agricultural products, the transportation environment can have a significant impact on the quality, freshness and shelf life of the agricultural products. For example, temperature fluctuations can accelerate the respiration of fruits and vegetables and the reproduction of microorganisms, and high humidity can easily cause mold. Therefore, the application real-time collects environmental data during the transportation of agricultural products during the logistics transportation process, and pre-processes the environmental data.

[0033] Specifically, the high-precision sensors deployed in the transportation environment real-time collect environmental data in the transportation environment, including but not limited to temperature, humidity, etc.

[0034] It should be noted that the noise interference can be avoided by denoising the environmental data. Therefore, in the embodiment, Kalman filtering is used to denoise the environmental data. In other embodiments, the implementer can select the method for denoising the environmental data according to the actual implementation, for example, using wavelet transform to denoise the environmental data.

[0035] The image of the surface of the agricultural product is collected in real time by using the camera deployed in the transportation environment, and the collected image is converted into a gray value image for subsequent processing.

[0036] S2, according to the difference between the agricultural product gray value images at the current time and the initial time, the state index of the agricultural product at the current time is extracted.

[0037] It should be noted that, in order to realize dynamic monitoring of the state change of the agricultural product in the transportation process, the present application is based on the sequence of continuously collected agricultural product gray value images, and the key index reflecting the freshness and physical state change of the agricultural product is extracted by image analysis technology.

[0038] Specifically, for the agricultural product gray value image at each time, the agricultural product gray value image is instance segmented to identify each independent agricultural product individual in the agricultural product gray value image. In the embodiment, the Mask R-CNN instance segmentation model is used to process the agricultural product gray value image, and the pixel-level segmentation mask and the boundary box of each agricultural product are output, so as to realize the separation and positioning of the agricultural product individuals.

[0039] It should be noted that, since the agricultural product may be locally displaced, rotated or occluded during transportation, in order to establish the correspondence relationship of individuals across time sequences, the present application further adopts a multi-target cross-frame matching strategy to match each agricultural product individual in the agricultural product gray value image at the current time with the agricultural product in the agricultural product gray value image at the initial time of transportation.

[0040] In the embodiment, the cross-frame matching of the agricultural product individuals is realized by using the method combining the SIFT key point matching and the Hungarian algorithm, and the specific process is as follows: The SIFT feature points of each agricultural product region in the agricultural product gray value images at the initial time of transportation and the current time are extracted, the feature points between each agricultural product region in the agricultural product gray value image at the initial time of transportation and each agricultural product region in the agricultural product gray value image at the current time are compared by using the nearest neighbor distance to preliminarily match, a similarity matrix between the agricultural product regions in the agricultural product gray value image at the initial time of transportation and the agricultural product regions in the agricultural product gray value image at the current time is constructed, the global optimal matching solution of the similarity matrix is solved by using the Hungarian algorithm, and the matching relationship of the agricultural product regions in the agricultural product gray value images at the initial time of transportation and the current time is obtained.

[0041] It should be noted that during the transportation of agricultural products, factors such as temperature fluctuations, humidity imbalances, or long transportation times may cause them to lose water, wilt, suffer cell structure damage, or partially rot, leading to significant changes in their appearance and surface features. These changes manifest themselves in images as contour shrinkage, surface wrinkling, browning, blurred texture, or localized mold, and are important visual indicators for measuring the decline in freshness of agricultural products. Therefore, the present invention comprehensively measures the degree of degradation of agricultural products during transportation using multiple quantifiable visual dimensions, such as the rate of change in area, grayscale change, and texture change, providing an objective basis for dynamic freshness assessment.

[0042] Specifically, the agricultural product area change rate is determined based on the area difference between each agricultural product area in the current agricultural product grayscale image and the corresponding agricultural product area in the agricultural product grayscale image at the initial transportation moment:

[0043] Among them, the current time sequence number is used To express, It indicates the rate of change of agricultural product area at the current moment, reflecting the average morphological degradation level of agricultural product groups; Indicates the current moment of the agricultural product grayscale image the area of ​​agricultural product areas; Indicates the current moment of the agricultural product grayscale image The area of ​​the agricultural product region corresponding to the agricultural product grayscale image at the initial moment of transportation; Indicates the number of effectively matched agricultural product regions in the current agricultural product grayscale image. Effective matching means that a one-to-one correspondence can be successfully established between the agricultural product regions in the current agricultural product grayscale image and the agricultural product regions in the agricultural product grayscale image at the initial transportation moment. Indicates the maximum value symbol, which is used to ensure that only the case of area reduction is counted, and to exclude misjudgments caused by changes in viewing angle or slight expansion. When the area change rate is larger, it means that the overall water loss of the agricultural product is more serious, the cell turgor pressure decreases, the degree of tissue wilting is higher, and the freshness is significantly reduced. For example, the curling of leafy vegetables and the wrinkling of the surface of fruits and vegetables will cause the segmented area to decrease, which will be reflected as a higher grayscale image of the agricultural product at the current moment. value.

[0044] Furthermore, the grayscale change rate of the agricultural product is determined based on the grayscale difference between each agricultural product area in the current agricultural product grayscale image and the corresponding agricultural product area in the agricultural product grayscale image at the initial transportation moment:

[0045] in, a serial number representing a current time, a gray scale change rate of the agricultural product at the current time; an average gray scale value of the agricultural product region No. in the gray scale image of the agricultural product at the current time; an average gray scale value of the agricultural product region corresponding to the agricultural product region No. in the gray scale image of the agricultural product at the initial time of transportation in the gray scale image of the agricultural product at the current time; a number of effectively matched agricultural product regions in the gray scale image of the agricultural product at the current time; an absolute value symbol, a maximum value symbol, a maximum value in the and and is obtained, which is used to prevent the denominator from being 0. The gray scale change rate reflects the evolution trend of the color depth and glossiness of the surface of the agricultural product. In a fresh state, the color of the surface of most agricultural products is uniform, and the reflection is moderate, which corresponds to a stable gray scale distribution; as the freshness decreases, browning, yellowing, mold spots or surface water may occur, which causes the local gray scale value to significantly increase (such as the whitening of moldy areas) or decrease (such as the darkening of rotten areas). The present application can effectively capture such degradation characteristics by calculating the relative change of the average gray scale. For example, enzymatic browning occurs in leafy vegetables during the oxidation process, and the gray scale value decreases significantly; and the gray scale value of some fruits and vegetables may increase due to the formation of a reflective film on the surface due to water loss or mold.

[0046] Further, according to the texture difference between each agricultural product region in the gray scale image of the agricultural product at the current time and the corresponding agricultural product region in the gray scale image of the agricultural product at the initial time of transportation, a texture change rate of the agricultural product is determined:

[0047] wherein, a serial number representing a current time, a texture change rate of the agricultural product at the current time; an entropy value of the gray scale co-occurrence matrix of the agricultural product region No. at the current time; an entropy value of the gray scale co-occurrence matrix of the agricultural product region corresponding to the agricultural product region No. in the gray scale image of the agricultural product at the current time in the gray scale image of the agricultural product at the initial time of transportation; a number of effectively matched agricultural product regions in the gray scale image of the agricultural product at the current time; an absolute value symbol, a maximum value symbol, a maximum value in the and and is obtained, which is used to prevent the denominator from being 0.

[0048] The entropy value of the gray level co-occurrence matrix is obtained as follows: the gray level co-occurrence matrix of each agricultural product region in the agricultural product gray image at the current time and the agricultural product gray image at the initial time of transportation in different directions (0°, 45°, 90°, 135°) is obtained respectively, the entropy value is calculated according to the gray level co-occurrence matrix in each direction, and the mean value of the entropy values corresponding to the gray level co-occurrence matrix of the agricultural product region in all directions is taken as the final result, so that the direction deviation is eliminated.

[0049] The entropy value of the gray level co-occurrence matrix reflects the complexity and irregularity of the gray distribution in the corresponding agricultural product region, and the greater the value, the more chaotic the texture. When the texture change rate is greater, it indicates that the surface texture of the agricultural product is more irregular, the local gray distribution is more chaotic, and it reflects that the quality deterioration phenomena such as mildew, rot, water or significant shrinkage may have occurred. For example, at the initial stage of gray mold of strawberries, although there is no obvious color change on the surface, fine velvet structure is formed, which leads to a significant increase in the entropy value; during the wilting process of leafy vegetables, the leaf surface wrinkles increase, which also causes the increase of the entropy value.

[0050] Further, according to the agricultural product area change rate, the agricultural product gray level change rate and the agricultural product texture change rate, the agricultural product state characteristic index is obtained:

[0051] wherein, represents the serial number at the current time, represents the agricultural product state characteristic index at the current time, which is used to represent the freshness level of the agricultural product; represents the agricultural product area change rate at the current time, which reflects the degree of morphological contraction of the agricultural product caused by water loss, wilting and the like; represents the agricultural product gray level change rate at the current time, which reflects the color degradation phenomena such as browning and discoloration on the surface of the agricultural product; represents the agricultural product texture change rate at the current time, which reflects the quality deterioration phenomena such as shrinkage, mold spot and rot on the surface of the agricultural product; represents the natural exponential function, which is used for negative correlation normalization of the agricultural product area change rate, the agricultural product gray level change rate and the agricultural product texture change rate, realizes mapping of the agricultural product area change rate, the agricultural product gray level change rate and the agricultural product texture change rate between [0, 1], eliminates the deviation caused by different physical dimensions and numerical ranges, and ensures the comparability and equal weight of each index when fused.

[0052] When the agricultural product area change rate , the agricultural product gray level change rate or the agricultural product texture change rate The larger, the more obvious degradation of the agricultural products in any of the shape, color or texture, and the lower the freshness level of the agricultural products. Conversely, if the area change rate of the agricultural products , the gray scale change rate of the agricultural products or the texture change rate of the agricultural products is small, the state characteristic index of the agricultural products is maintained at a high value, indicating that the state of the agricultural products is stable and the quality is good during transportation.

[0053] S3, based on the particle swarm optimization algorithm, the state risk value of the transportation scheme corresponding to each particle is determined combined with the predicted environmental data and the state index of the agricultural products at each time of the whole transportation process; the inertia weight and the social learning factor of each particle are dynamically adjusted according to the state risk value, and the cold chain transportation scheme including the optimal driving speed, the set temperature and humidity and the distribution path is output by iterative optimization taking the fitness function as the target.

[0054] It should be noted that during the cold chain transportation of agricultural products, the dynamic change of the transportation environment and the degradation process of the quality of the agricultural products are coupled with each other, and the traditional particle swarm optimization algorithm usually uses a fixed or globally unified inertia weight, which is difficult to dynamically adjust the search strategy according to the current transportation state, resulting in rough search in the high-risk stage, slow convergence in the stable stage, and unable to achieve the optimal balance between energy consumption and preservation. Therefore, the present application proposes a state-aware adaptive inertia weight field mechanism, which integrates the state index of the agricultural products and the environmental data into a state risk signal for dynamically adjusting the inertia weight and the social learning factor of each particle in the particle swarm optimization algorithm, so that the optimization process can automatically switch between the "exploration" and "development" behaviors according to the transportation risk level, and improve the search efficiency and robustness of the algorithm in complex dynamic environment.

[0055] The flow chart of step S3 is referred to Figure 2 , including steps S301-S305: S301, for each particle in each particle swarm optimization iteration process, based on the transportation scheme represented by the particle, combined with real-time environmental data and historical transportation state, the evolution trend of the state of the agricultural products under the transportation scheme is predicted.

[0056] Specifically, the transportation scheme represented by the i-th particle in the j-th iteration is represented as , wherein represents the driving speed, represents the set temperature, represents the set humidity, represents the distribution path. The following prediction steps are performed based on the transportation scheme

[0057] The following prediction steps are performed based on the transportation scheme ​​1. According to the transportation path and the driving speed , combined with the distance of each section, the position of the vehicle at each time is calculated.

[0058] 2. Based on the historical meteorological data of each area on the transportation path, the temperature and humidity inside the carriage of the transportation vehicle at each time are predicted.

[0059] In one embodiment, the prediction method of temperature at each time is as follows: The outdoor temperature of each section on the transportation path is obtained through the public meteorological platform, the vehicle insulation coefficient is obtained according to the technical parameters of the vehicle, for example, the insulation coefficient of a class A insulated vehicle is 0.3, that is, for every 1℃ change in outdoor temperature, the temperature of the carriage is affected by a fluctuation of 0.3℃; and the rated cooling precision of the refrigeration system is obtained.

[0060] According to the predicted position of the vehicle at each time, the outdoor temperature of the section corresponding to the position is subtracted from the set temperature of the vehicle to obtain the deviation value of the outdoor temperature and the set temperature ; the deviation value is multiplied by the vehicle insulation coefficient to obtain the influence of the outdoor environment on the temperature of the carriage; the set temperature is superimposed with the influence, and then corrected in combination with the rated cooling precision of the refrigeration system to obtain the predicted value of the internal temperature of the carriage at the corresponding time.

[0061] In one embodiment, the prediction method of humidity at each time is as follows: The humidity data of each section on the transportation path is obtained through the public meteorological platform, the air exchange rate between the carriage and the outside world is obtained according to the sealing performance parameters of the vehicle, for example, the air exchange rate of a common sealed carriage is 10%, that is, the carriage exchanges 10% of the air with the outside world every hour; the transpiration and humidity increase per unit time is obtained according to the type of the transported agricultural products, for example, leafy vegetables increase by 0.3% per 100kg per 10 minutes, and root vegetables increase by 0.5% per 100kg per 10 minutes, which is pre-calibrated through transportation experiments of the same type of agricultural products; and the rated dehumidification efficiency of the dehumidification system is obtained, for example, the humidity can be reduced by 0.5% every 10 minutes.

[0062] The outdoor humidity per hour in the section is divided into each time according to the air exchange rate to obtain the humidity increment brought in by the outside world at that time, for example, the air exchange every 10 minutes brings in the outside humidity of . According to the predicted position of the vehicle at each time, the humidity data of the section corresponding to the position is multiplied by the air exchange rate to obtain the humidity increment brought into the carriage by the outside air; the humidity increment is summed with the transpiration and humidity increase of the agricultural products to obtain the total increase of the humidity of the carriage; and the set humidity of the vehicle It is added to the total increase and then the difference is calculated with the rated dehumidification efficiency of the dehumidification system to obtain the predicted value of the humidity inside the car at the corresponding moment.

[0063] 3. Use Long Short-Term Memory Network (LSTM) to predict transportation routes Evolutionary trends of agricultural product status indicators: The LSTM network takes as input a sequence of environmental data and a sequence of historical agricultural product status indicators, and outputs predicted values ​​for agricultural product status indicators at each future moment from the current moment to the transport destination. The environmental data sequence consists of two parts: historical temperature and humidity data from the start of transport to the current moment (collected in real time by sensors); and a sequence of predicted temperature and humidity data from the current moment to the transport destination. These two are concatenated in chronological order to form a complete temperature and humidity sequence. The historical agricultural product status indicator sequence specifically refers to the agricultural product status indicators extracted from the start of transport to the current moment.

[0064] S302: Calculate the state risk value of the transportation plan corresponding to each particle based on the predicted temperature, humidity, and change rate of the agricultural product state characteristic indicators.

[0065] It's important to note that environmental parameters or status indicators at a single moment in time cannot accurately reflect the overall risk level of the transportation process, and the deterioration of agricultural product quality is the cumulative effect of environmental factors. Therefore, the present invention calculates the changing trends of agricultural product status characteristic indicators under transportation plans and, combined with the environmental deviation of the entire transportation process, comprehensively assesses the risk level of the transportation plan.

[0066] Specifically, the deterioration rate of the agricultural product status characteristic indicators is calculated:

[0067] in, Indicates the The degradation rate of the characteristic index of the agricultural product status of the transportation scheme represented by each particle reflects the speed of the decline in the quality of agricultural products. Indicates the The particle represents the predicted characteristic index of the agricultural product status when arriving at the transportation destination under the transportation plan; Indicates the characteristic indicators of agricultural product status at the current moment; Indicates the remaining time from the current moment to the destination of transportation. When it is large, it indicates that the quality of agricultural products is deteriorating rapidly and the transportation plan faces a high risk. On the contrary, it indicates that the environmental control under the transportation plan is effective and the quality is maintained in good condition. The maximum function is used to ensure that the deterioration rate is non-negative. When the predicted state characteristic index of the agricultural products at the transportation end point is better than that at the current time, it indicates that the quality of the agricultural products does not deteriorate or even remains stable due to the optimized environment under the transportation scheme. At this time, the deterioration rate is forced to be zero to avoid interference of the reverse value on the risk assessment. Only when the predicted state characteristic index of the agricultural products at the transportation end point is lower than that at the current time, the deterioration rate is calculated according to the actual difference to truly reflect the speed of quality decline, and ensure the objectivity and rationality of the state risk value calculation.

[0068] Further, the temperature deviation degree of the entire transportation process is constructed. The deviation of the predicted temperature of each time under the transportation scheme represented by the particle from the optimal storage temperature interval of the transported agricultural products is counted. For the predicted temperature of any time, if the predicted temperature is lower than the lower limit of the optimal storage temperature interval, the ratio of the difference between the predicted temperature and the lower limit to the total length of the optimal storage temperature interval is taken as the deviation ratio of the predicted temperature. If the predicted temperature is higher than the upper limit of the optimal storage temperature interval, the ratio of the difference between the predicted temperature and the upper limit to the total length of the optimal storage temperature interval is taken as the deviation ratio of the predicted temperature. If the predicted temperature is within the optimal storage temperature interval, the deviation ratio of the predicted temperature is 0. The normalized result of the mean of the deviation ratios of the predicted temperatures of all times under the transportation scheme is taken as the temperature deviation degree.

[0069] The greater the temperature deviation degree is, the more ideal the temperature control is, and the higher the threat to the quality of the agricultural products is.

[0070] Similarly, the humidity deviation degree of the entire transportation process is constructed. The deviation of the predicted humidity of each time under the transportation scheme represented by the particle from the optimal storage humidity interval of the transported agricultural products is counted. For the predicted humidity of any time, if the predicted humidity is lower than the lower limit of the optimal storage humidity interval, the ratio of the difference between the predicted humidity and the lower limit to the total length of the optimal storage humidity interval is taken as the deviation ratio of the predicted humidity. If the predicted humidity is higher than the upper limit of the optimal storage humidity interval, the ratio of the difference between the predicted humidity and the upper limit to the total length of the optimal storage humidity interval is taken as the deviation ratio of the predicted humidity. If the predicted humidity is within the optimal storage humidity interval, the deviation ratio of the predicted humidity is 0. The normalized result of the mean of the deviation ratios of the predicted humidities of all times under the transportation scheme is taken as the humidity deviation degree.

[0071] The greater the humidity deviation degree is, the more ideal the humidity control is, and the higher the threat to the quality of the agricultural products is.

[0072] It should be noted that the optimal storage temperature range and the optimal storage humidity range are set according to the type of agricultural products and the transportation and preservation requirements. For example, the optimal storage temperature range of leafy vegetables is usually 0℃-4℃, and the optimal storage humidity range is 95%-100%; the optimal storage temperature range of root vegetables is usually 5℃-10℃, and the optimal storage humidity range is 70%-80%. The optimal storage temperature range and the optimal storage humidity range of different agricultural products can be pre-set based on the research data of agricultural preservation, industry standards or historical transportation experimental results, and can be adjusted in detail according to specific categories (such as spinach, potatoes, strawberries, etc.).

[0073] Further, according to the deterioration rate of the state characteristic index of the agricultural products represented by the particle and the temperature deviation degree and the humidity deviation degree of the whole transportation process, the state risk value of the particle is determined:

[0074] wherein, represents the state risk value of the i-th particle at the j-th iteration, which is used to represent the comprehensive deterioration pressure of the transportation scheme represented by the particle under the current environment; represents the deterioration rate of the state characteristic index of the agricultural products under the transportation scheme within the sliding time window; and are the temperature deviation degree and the humidity deviation degree of the transportation scheme represented by the i-th particle, respectively; is the temperature deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the temperature deviation degree of the transportation scheme represented by the i-th particle and the temperature deviation degree of the transportation scheme represented by the j-th particle; is the humidity deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the humidity deviation degree of the transportation scheme represented by the i-th particle and the humidity deviation degree of the transportation scheme represented by the j-th particle; is the temperature deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the temperature deviation degree of the transportation scheme represented by the i-th particle and the temperature deviation degree of the transportation scheme represented by the j-th particle; is the temperature deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the temperature deviation degree of the transportation scheme represented by the i-th particle and the temperature deviation degree of the transportation scheme represented by the j-th particle; is the humidity deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the humidity deviation degree of the transportation scheme represented by the i-th particle and the humidity deviation degree of the transportation scheme represented by the j-th particle; is the humidity deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the humidity deviation degree of the transportation scheme represented by the i-th particle and the humidity deviation degree of the transportation scheme represented by the j-th particle; is the humidity deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the humidity deviation degree of the transportation scheme represented by the i-th particle and the humidity deviation degree of the transportation scheme represented by the j-th particle; is the humidity deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the humidity deviation degree of the transportation scheme represented by the i-th particle and the humidity deviation degree of the transportation scheme represented by the j-th particle; is the humidity deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the humidity deviation degree of the transportation scheme represented by the i-th particle and the humidity deviation degree of the transportation scheme represented by the j-th particle; is the humidity deviation degree of the transportation scheme represented by the i-th particle, which is calculated according to the humidity deviation degree of the transportation scheme represented by the i-th particle and the humidity deviation degree of the transportation scheme represented by the j-th particle;

[0075] S303, dynamically adjusting the inertia weight and the social learning factor of each particle according to the state risk value.

[0076] It should be noted that the inertia weight is a parameter that controls the inertia of particles maintaining their previous velocity, directly affecting the algorithm's global exploration and local exploitation capabilities. To enable the optimization process to automatically adjust search behavior based on the level of transportation risk, the present invention dynamically adjusts the inertia weight of each particle based on its state risk value, automatically enabling high-risk solutions to enter a refined search mode.

[0077] Specifically, the inertia weight of each particle satisfies the expression:

[0078] in, Indicates the The particle in Inertia weight at iteration ; and They are the lower and upper limits of the preset inertia weight, which are used to limit the scope of the search behavior. The empirical value is , ; Indicates the The particle in The state risk value at the iteration; is the attenuation coefficient, which is used to control the response sensitivity of the inertia weight to the change of risk. The empirical value is When the state risk value When smaller, Close to the upper limit of inertia weight , the particles maintain a large historical velocity inertia and conduct a large-scale exploration; when the state risk value When it is larger, Approaching the lower limit of inertia weight , the particles enter the fine search mode to avoid missing the optimal temperature control or path strategy due to too large a step size in high-risk areas.

[0079] It should be noted that the social learning factor is a parameter that controls the tendency of particles to approach the global optimal position, affecting the convergence speed and diversity of the algorithm. To prevent high-risk particles from blindly following the global optimal solution and falling into local convergence, the present invention dynamically adjusts the social learning factor.

[0080] Specifically, the social learning factor of each particle satisfies the expression:

[0081] in, Indicates the The particle in Social learning factor at the iteration; is the initial social learning factor, with an empirical value of 2.0; Indicates the The particle in state risk value at the current iteration; is a risk influence decay coefficient, used to control the degree of weakening of the social learning factor with the increase of risk, and the experience value is . When the state risk value is larger, the social learning factor is smaller, and the particle reduces the dependence on the global optimal position, enhancing the individual autonomous exploration ability. Conversely, when the state risk value is lower, the social learning factor is close to the initial value, and the particle is more inclined to follow the group optimal solution, accelerating the convergence speed.

[0082] S304, updating the speed and position of the particle group according to the inertia weight and the social learning factor of the particle.

[0083] Specifically, the speed and position of the particle at each iteration satisfy the expression:

[0084]

[0085] wherein, and respectively represent the speed and position of the i-th particle at the j-th iteration; and respectively represent the speed and position of the i-th particle at the j-th iteration; is the individual historical optimal position of the i-th particle; is the current global optimal position; is an individual learning factor, which is 2.0 in this embodiment, and in other embodiments, the implementer can set according to the actual implementation situation; represents the inertia weight of the i-th particle at the j-th iteration; represents the social learning factor of the i-th particle at the j-th iteration; , is a random number in the range of [0, 1]. , is a random number in the range of [0, 1].

[0086] S305, constructing an adaptive function, and performing iterative optimization with the adaptive function as the target, to output a cold chain transportation scheme containing the optimal driving speed, the set temperature and humidity, and the distribution path.

[0087] Specifically, the adaptive function satisfies the expression:​​​​​​​

[0088] in, Indicates the The fitness value of each particle; Indicates the Total energy consumption for cooling and dehumidification under the transportation scheme corresponding to each particle; Indicates the The predicted characteristic indicators of agricultural product status when arriving at the transportation destination under the transportation plan corresponding to each particle; and Respectively The temperature deviation and humidity deviation of the transportation scheme represented by each particle; 、 、 is the weight coefficient. In this embodiment, we pay more attention to the quality of agricultural products, so the weight coefficient is set to 、 、 , implementation personnel can also set it according to the actual implementation situation, but must ensure 、 、 The sum is 1.

[0089] Repeat steps S301 to S305 until the maximum number of iterations is reached. Or convergence conditions, output the global optimal solution The global optimal solution includes the optimal driving speed, set temperature and humidity, distribution path and other parameters, which constitute the final cold chain logistics optimization plan. Among them, the maximum number of iterations The convergence conditions are set by the implementer according to the actual implementation situation. In this embodiment, , the convergence condition is: continuous The global optimal solution in the iteration The change in fitness value is less than the preset threshold .

[0090] It should be noted that this invention transforms the optimization process from "static parameter control" to "state-driven adaptive search" by fusing agricultural product status indicators with environmental data into status risk signals, which are then used to dynamically regulate the inertia weight and social learning factor of each particle in the particle swarm optimization algorithm. This mechanism automatically adjusts search behavior based on the level of transportation risk, improving search accuracy in high-risk phases and accelerating convergence in low-risk phases, significantly enhancing the intelligent level of cold chain logistics optimization.

[0091] The embodiment of the present application also discloses an in-transit monitoring management system for logistics transportation, comprising a processor and a memory, and the memory stores computer program instructions which realize the in-transit monitoring management method for logistics transportation according to the present application when executed by the processor.

[0092] The above system also comprises other components such as a communication bus and a communication interface which are well known to those skilled in the art, and the setting and functions thereof are known in the art, thus not described here again.

Claims

1. A method for in-transit monitoring and management of logistics transportation, characterized in that: include: Collect transportation environment data and grayscale images of agricultural products in real time, and extract agricultural product status indicators based on the grayscale images; Based on the particle swarm optimization algorithm, the state risk value of the transportation plan corresponding to each particle is determined by combining the predicted environmental data and agricultural product status indicators at each moment of the transportation process. The inertia weight and social learning factor of each particle are dynamically adjusted according to the state risk value, and iterative optimization is performed with the fitness function as the target. The output is a cold chain transportation plan that includes the optimal driving speed, set temperature and humidity, and distribution route.

2. The method for in-transit monitoring and management of logistics transportation according to claim 1, characterized in that: The step of extracting agricultural product status indicators based on grayscale images includes: The grayscale image is segmented and the agricultural product area is extracted. The agricultural product area at the current moment and the initial moment are matched through SIFT key point matching and Hungarian algorithm. The area change rate, grayscale change rate and texture change rate of the agricultural product area between grayscale images of different frames are calculated, and the agricultural product status characteristic indicators are obtained by fusion.

3. The method for in-transit monitoring and management of logistics transportation according to claim 2, characterized in that: The area change rate is the average of the area shrinkage ratios of the corresponding agricultural product areas at the current moment and the initial moment; The grayscale change rate is the average of the absolute values ​​of the relative changes in the average grayscale of the agricultural product area corresponding to the current moment and the initial moment; The texture change rate is the absolute value mean of the relative change of the gray level co-occurrence matrix entropy value of the agricultural product area corresponding to the current moment and the initial moment.

4. The method for in-transit monitoring and management of logistics transportation according to claim 1, characterized in that: The method for obtaining the state risk value is: According to the transportation plan represented by the particle, the difference between the predicted agricultural product status characteristic index at the transportation destination and the current status characteristic index is divided by the remaining time from the current moment to the transportation destination to obtain the degradation rate of the agricultural product status characteristic index; the state risk value is obtained by integrating the degradation rate, the temperature deviation, and the humidity deviation of the entire transportation process of the transportation plan.

5. The method for in-transit monitoring and management of logistics transportation according to claim 4, characterized in that: The temperature deviation is the normalized result of the mean of the deviation ratio of the predicted temperature at each moment compared to the optimal storage temperature range of the agricultural product under the transportation scheme represented by the particle; The humidity deviation is the normalized result of the mean of the deviation ratios of the predicted humidity at each moment compared to the optimal storage humidity range of the agricultural product under the transportation scheme represented by the particle.

6. The method for in-transit monitoring and management of logistics transportation according to claim 4, characterized in that: The state risk value satisfies the expression: ； in, Indicates the The particle in The state risk value at the iteration; Indicates the The degradation rate of the agricultural product status characteristic index within the sliding time window under the transportation scheme represented by each particle; and Respectively The temperature deviation and humidity deviation of the transportation scheme represented by each particle; is the sensitivity weight coefficient of agricultural product type.

7. The method for in-transit monitoring and management of logistics transportation according to claim 1, characterized in that: The inertia weight satisfies the expression: ； in, Indicates the The particle in Inertia weight at iteration ; and are the lower and upper limits of the preset inertia weight respectively; Indicates the The particle in The state risk value at the iteration; is the attenuation coefficient.

8. The method for in-transit monitoring and management of logistics transportation according to claim 1, characterized in that: The social learning factor satisfies the expression: ； in, Indicates the The particle in Social learning factor at the iteration; is the initial social learning factor; Indicates the The particle in The state risk value at the iteration; is the risk influence attenuation coefficient.

9. The method for in-transit monitoring and management of logistics transportation according to claim 1, characterized in that: The fitness function satisfies the expression: ； in, Indicates the The comprehensive fitness value of each particle; Indicates the Total energy consumption for cooling and dehumidification under the transportation scheme corresponding to each particle; Indicates the The predicted characteristic indicators of agricultural product status when arriving at the transportation destination under the transportation plan corresponding to each particle; and Respectively The temperature deviation and humidity deviation of the transportation scheme represented by each particle; 、 、 is the weight coefficient.

10. An in-transit monitoring and management system for logistics transportation, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an in-transit monitoring and management method for logistics transportation according to any one of claims 1 to 9 is implemented.

Images