Multi-mode AIGC cold chain cargo state monitoring method and system

Abstract

The invention relates to the field of cold chain cargo state monitoring, in particular to a multi-mode AIGC cold chain cargo state monitoring method and system. The method comprises the following steps: acquiring a real-time infrared monitoring image of the cold-chain goods; performing target visual identification on the real-time infrared monitoring image of the cold-chain goods, fitting temperature distribution of the goods in the region, and constructing a temperature distribution diagram of the goods in the region; acquiring an omnibearing conventional monitoring image in the cold-chain cargo transportation warehouse; performing in-warehouse three-dimensional layout distribution analysis on the omni-directional conventional monitoring image according to a cargo temperature distribution diagram in the region, and performing dynamic temperature distribution rendering to obtain a dynamic temperature distribution twin model; and carrying out temperature change fluctuation amplitude calculation and multi-time-point temperature change situation prediction on the dynamic temperature distribution twin model so as to generate a multi-time-point temperature change situation prediction map of each sub-region. According to the invention, the real-time and accurate evaluation of the cargo state is realized, and the intelligent management decision is provided.

Description

Technical Field

[0001] The present invention relates to the field of cold chain cargo status monitoring, and in particular to a multi-modal AIGC cold chain cargo status monitoring method and system. Background Art

[0002] With the development of the global economy and the diversification of consumer demand, cold chain logistics has become an important link in ensuring the quality and safety of high-value goods such as fresh food, medicines, and chemicals. The core of cold chain logistics is to ensure that the goods can always be in a temperature and humidity environment that meets the requirements during the entire transportation and storage process to prevent the goods from deteriorating, being damaged or failing due to temperature control failure or environmental changes. However, traditional cold chain monitoring systems face many challenges in practical applications, such as single monitoring methods, low frequency of data collection, frequent manual intervention and errors, and difficulty in achieving full-process dynamic monitoring.

[0003] At present, the cold chain logistics monitoring system mainly relies on temperature and humidity sensors for single environmental data collection. Such traditional sensors can usually only provide limited temperature and humidity information and cannot fully reflect the changes in the complex transportation environment. For example, changes in temperature and humidity are affected by a variety of factors such as the location of the goods, the driving status of the freight vehicles, and external weather changes. It is difficult to fully evaluate the status of the goods only by relying on temperature and humidity data. In addition, traditional monitoring methods often rely on manual inspections and regular inspections, which are prone to omissions or delays and cannot reflect the actual status of the goods in real time and accurately.

[0004] With the development of multimodal AIGC (AI-Generated Content), AIGC refers to artificial intelligence technology that can generate content in various media forms. It is an important development direction in the field of artificial intelligence in recent years. Multimodal data refers to information from multiple aspects such as environment, cargo status, and transportation equipment collected by different types of sensors and monitoring equipment. For example, temperature sensors, humidity sensors, CO2 concentration sensors, airflow sensors, and video surveillance cameras work together to obtain more comprehensive and accurate environmental information through data fusion. By combining these multimodal data with advanced artificial intelligence algorithms, the system can monitor and intelligently analyze the status of the cargo in real time, thereby achieving early warning of potential risks. Therefore, the cold chain cargo monitoring method based on multimodal data fusion has gradually become an important trend to improve the intelligence level of the cold chain monitoring system. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes a multimodal AIGC cold chain cargo status monitoring method and system to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a multi-modal AIGC cold chain cargo status monitoring method, comprising the following steps: Step S1: Acquire a real-time infrared monitoring image of cold chain goods; perform target visual recognition on the real-time infrared monitoring image of cold chain goods, perform temperature distribution fitting on the goods in the area, and construct a temperature distribution map of the goods in the area; Step S2: Obtain a full-range conventional monitoring image in the cold chain cargo transport warehouse; perform a three-dimensional layout distribution analysis in the warehouse on the full-range conventional monitoring image according to the cargo temperature distribution map in the area, and perform dynamic temperature distribution rendering to obtain a dynamic temperature distribution twin model; Step S3: Calculate the temperature fluctuation amplitude and predict the temperature change trend at multiple time points for the dynamic temperature distribution twin model, thereby generating a multi-time point temperature change trend prediction map for each sub-region; Step S4: mining abnormal temperature fluctuations on the multi-time point temperature change trend prediction map of each sub-region, and tracing the abnormal region to obtain the abnormal temperature change cargo node; Step S5: Analyze the surface texture details based on the abnormal temperature change cargo node, and then perform packaging damage detection to generate abnormal temperature change cargo damage detection results; Step S6: Make abnormal temperature emergency decisions and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results, and build an intelligent cargo status decision model.

[0007] The present invention obtains real-time infrared monitoring images of cold chain goods and performs target visual recognition to accurately identify goods and obtain real-time temperature data. Constructing a temperature distribution map of goods in the area helps to understand the temperature distribution of goods and discover temperature anomalies or hot spots in advance, thereby ensuring the safety of cold chain transportation of goods. Dynamic temperature distribution rendering is combined with three-dimensional layout distribution analysis in the warehouse to generate a dynamic temperature distribution twin model, providing all-round temperature monitoring and analysis. Through the dynamic temperature distribution twin model, real-time monitoring and prediction of temperature changes of goods in the warehouse are achieved, which helps to ensure the quality of goods. Multi-time point temperature change trend prediction helps to predict the changing trend of cargo temperature and make timely adjustments and response measures. Generating a multi-time point temperature change trend prediction map for each sub-area helps to discover temperature anomalies or hot spots in advance. Abnormal temperature conditions are detected to ensure that the goods are within a safe temperature range. Abnormal temperature fluctuation mining and abnormal area tracing help to accurately locate the cause and scope of abnormal temperature conditions. By generating abnormal temperature change cargo nodes, timely measures are taken to deal with abnormal conditions and ensure the quality of goods. The combination of surface texture detail analysis and package damage detection helps to identify whether the goods are damaged due to abnormal temperature. The abnormal temperature change cargo damage detection results are generated to provide a comprehensive assessment of the cargo status and provide a basis for subsequent processing. Abnormal temperature emergency decision-making and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results help to deal with abnormal conditions in a timely manner and ensure the safety of goods. The construction of an intelligent cargo status decision-making model improves decision-making efficiency and accuracy, and provides intelligent management and guarantee for cold chain cargo transportation.

[0008] Preferably, step S1 comprises the following steps: Step S11: Perform infrared imaging monitoring on the cold chain cargo transport warehouse to obtain real-time infrared monitoring images of the cold chain cargo; Step S12: Performing target visual recognition on the real-time infrared monitoring image of cold chain goods and marking each cargo node; Step S13: accurately locate the boundary of each cargo node in the region to generate the regional position coordinates of each cargo node; Step S14: performing temperature gradient distribution detection on the real-time infrared monitoring image of cold chain goods, thereby obtaining image temperature gradient distribution characteristics; Step S15: fitting the temperature distribution of cargo in the region to the regional position coordinates of each cargo node according to the temperature gradient distribution characteristics in the region, and constructing a temperature distribution map of cargo in the region.

[0009] The present invention obtains real-time infrared monitoring images of cold chain goods through infrared imaging monitoring, realizes non-contact monitoring of the temperature of goods, and captures the temperature information of goods in time. This helps to realize temperature monitoring of goods in a cold chain environment and improves the real-time monitoring capability of the state of goods during transportation. Through target visual identification and marking of each cargo node, the location and quantity of the goods are accurately identified, providing basic data for subsequent analysis and processing. This helps to distinguish and track the goods individually and improve the accuracy and reliability of the monitoring system. Accurate boundary positioning in the region is performed to generate the regional location coordinates of each cargo node, which helps to accurately grasp the location and range of each cargo. The regional location coordinates of the cargo nodes are generated to help establish the spatial distribution information of the cargo and provide basic data for subsequent temperature distribution fitting. The temperature gradient distribution characteristics of the image are obtained by temperature gradient distribution detection, revealing the temperature changes in different regions, and providing a basis for subsequent temperature distribution analysis. This helps to understand the temperature change law more carefully and provide more reliable data support for the accurate fitting of the temperature distribution of the cargo. According to the temperature gradient distribution characteristics, the regional location coordinates of each cargo node are fitted with the temperature distribution of the cargo in the region, and a temperature distribution map of the cargo in the region is constructed to more accurately understand the temperature distribution of the cargo. By constructing a temperature distribution map of the cargo in the area, a visual display of the temperature distribution of the cargo can be achieved, providing a basis for further temperature control and management.

[0010] Preferably, step S2 comprises the following steps: Step S21: obtaining a omnidirectional conventional monitoring image in the cold chain cargo transport warehouse; reconstructing the global details of the omnidirectional conventional monitoring image to construct a detailed reconstructed monitoring image; Step S22: segmenting the boundary of each cargo image one by one on the detailed reconstructed monitoring image according to the regional position coordinates of each cargo node, thereby obtaining a plurality of cargo boundary image frames; Step S23: performing a three-dimensional morphological structure analysis on the boundary image frames of the plurality of goods to generate a three-dimensional morphological structure feature of each of the goods; Step S24: performing a three-dimensional layout distribution analysis on the boundary image frames of the multiple goods in the warehouse to obtain three-dimensional layout distribution data of the goods; Step S25: Performing three-dimensional point cloud modeling of the transport warehouse for the three-dimensional morphological structure characteristics of each cargo according to the three-dimensional layout distribution data of the cargo, and constructing a three-dimensional cargo distribution model; Step S26: Render the three-dimensional cargo distribution model with dynamic temperature distribution according to the cargo temperature distribution map in the area to obtain a dynamic temperature distribution twin model.

[0011] The present invention acquires omnidirectional conventional monitoring images and reconstructs the global details of the images to improve the clarity and detail expression of the monitoring images, making the monitoring process more accurate and reliable. Constructing detailed reconstructed monitoring images helps to present the detailed information of the goods and the environment more clearly, and improves the monitoring system's ability to understand and analyze the status of cold chain goods. By segmenting the boundaries of the goods images one by one, the boundary image frames of multiple goods are obtained, and the position and boundaries of each goods are accurately determined, providing accurate data support for subsequent analysis and processing, which helps to segment and identify the goods individually, laying the foundation for subsequent morphological structure analysis and layout distribution analysis. The three-dimensional morphological structure characteristics of each cargo are generated through three-dimensional morphological structure analysis, which more comprehensively describes the morphological characteristics of the goods and provides more information for the comprehensive evaluation of the goods status. The three-dimensional morphological structure characteristic analysis helps to understand the changes in the morphological characteristics of the goods, and provides a more in-depth reference basis for subsequent transportation and management decisions. The internal three-dimensional layout distribution analysis obtains the three-dimensional layout distribution data of the goods, reveals the spatial distribution of the goods in the warehouse, and provides a reference for cargo management and transportation. This helps to better understand the spatial layout of the goods and provide optimization solutions and decision support for the transportation and storage of cold chain goods. A three-dimensional cargo distribution model is constructed through three-dimensional point cloud modeling of the transport warehouse to achieve modeling and visualization of the goods in three-dimensional space, providing a convenient tool for the spatial management of goods in the transport warehouse. The three-dimensional cargo distribution model helps to intuitively display the spatial layout of the goods and provide a more intuitive and comprehensive information display for the cold chain cargo status monitoring system. The dynamic temperature distribution twin model is obtained through dynamic temperature distribution rendering, which combines the temperature distribution of the goods with their three-dimensional morphological structure to achieve a comprehensive display and analysis of the cargo status. The dynamic temperature distribution twin model helps to fully understand the temperature distribution of the goods and provides a deeper perspective for the dynamic monitoring and analysis of the cold chain cargo status.

[0012] Preferably, the specific steps of step S21 are: Perform in-warehouse ray tracing processing on all-round conventional monitoring images to obtain the ray tracing trajectory of the transport warehouse; Performing light distribution identification on the light tracing trajectory of the transport warehouse to generate light distribution data of the transport warehouse; Perform local brightness detection based on the light distribution data of the transport warehouse and extract the local dark areas in the image; Adaptively enhance the brightness of the local dark areas in the image to construct an adaptive brightness enhanced monitoring image; Perform computer vision detection of cargo contours on adaptive brightness enhanced surveillance images to extract the contours of each cargo; Perform edge detail sharpening on the contour line of each product to obtain the detailed sharpened contour features of each product; Based on the detailed sharpened contour features of each cargo, the global details of the adaptive brightness enhanced surveillance image are reconstructed to construct a detailed reconstructed surveillance image.

[0013] The present invention analyzes the light propagation in the transport warehouse through ray tracing processing, helps understand the distribution and path of light in the warehouse, and provides data support for subsequent light distribution identification. Through light distribution identification, the light distribution in the transport warehouse can be understood, which helps to evaluate the lighting conditions and light brightness distribution in the warehouse, and provides a basis for subsequent local dimming detection. Through dimming detection, the local dimming areas in the image are identified to help find potential areas with insufficient light or abnormal cargo conditions. Adaptive brightness enhancement processing improves the visibility of local dimming areas in the image, making the monitoring image clearer and easier to analyze, which helps to accurately monitor the status of the cargo. Through computer vision to detect the contour of the cargo, the contour information of each cargo is accurately extracted, providing accurate data support for subsequent detail sharpening and global detail reconstruction. Edge detail sharpening highlights the contour features of each cargo, enhances the clarity of the cargo edge in the image, and helps to more accurately analyze the shape and status of the cargo. Through global detail reconstruction of the image, the detailed features and overall visual effects of the cargo are comprehensively considered, the quality and analyzability of the monitoring image are improved, and a more accurate data basis is provided for cold chain cargo status monitoring.

[0014] Preferably, step S3 specifically comprises the following steps: Step S31: performing sub-region spatial division on the dynamic temperature distribution twin model, thereby obtaining a plurality of dynamic temperature sub-regions; Step S32: analyzing the time series variation of the temperature distribution of multiple dynamic temperature sub-regions to obtain the time series variation characteristics of the temperature distribution of each sub-region; Step S33: calculating the temperature change fluctuation amplitude of each sub-region temperature distribution time series change characteristic to generate the temperature change fluctuation amplitude characteristic of each sub-region; Step S34: Evolving the temperature change trend based on the temperature change fluctuation amplitude characteristics, and constructing a temperature change trend evolution curve for each sub-region; Step S35: Perform multi-time point temperature change situation prediction on the temperature change trend evolution curve of each sub-region, so as to generate a multi-time point temperature change situation prediction map of each sub-region. The present invention analyzes the temperature changes in the cold chain environment in more detail through sub-region spatial division, and subdivides the temperature distribution into multiple sub-regions, which is helpful for fine monitoring of the temperature status of different regions. The time-series change analysis of the temperature distribution reveals the temperature change trend of each sub-region, helps to understand the periodic and trend changes of temperature, and provides a basis for the subsequent calculation of the temperature change fluctuation amplitude. The temperature change fluctuation amplitude calculation quantifies the range of temperature changes in the sub-region, helps to evaluate the intensity of temperature fluctuations, and provides a quantitative reference for the subsequent evolution of temperature change trends. The temperature change trend evolution curve reflects the overall trend of temperature changes in the sub-region, helps to understand the long-term evolution law of temperature, and helps to predict future temperature change trends. The multi-time point temperature change situation prediction provides speculation on the temperature status of the sub-region at different time points, helps to monitor and predict the temperature status changes of goods in different time periods, and provides more dimensional information support for cold chain goods status monitoring.

[0015] Preferably, the specific steps of step S4 are: Step S41: performing temperature change mutation analysis on the multi-time point temperature change trend prediction map of each sub-region to identify the temperature change mutation time point; Step S42: mining abnormal temperature fluctuations at the temperature change mutation time point, thereby generating abnormal temperature fluctuation features; Step S43: tracing the source of the abnormal region according to the abnormal temperature fluctuation characteristics, thereby extracting the abnormal temperature change sub-region; Step S44: performing abnormal cargo node analysis on the abnormal temperature change sub-area, thereby obtaining an abnormal temperature change cargo node.

[0016] The present invention can timely identify the mutation time point of temperature change by performing temperature change mutation analysis on the multi-time point temperature change situation prediction map, improve the sensitivity and accuracy of the change of the state of cold chain goods, identify the mutation time point of temperature change, help to find the key moment of temperature anomaly, and provide important clues for subsequent abnormal detection and processing, generate abnormal temperature fluctuation characteristics through abnormal temperature fluctuation mining, more accurately capture and describe abnormal temperature fluctuation conditions, and provide a basis for abnormal detection and analysis, abnormal temperature fluctuation characteristics help to identify and quantify abnormal conditions, and provide important information for further abnormal area tracing and processing, trace the abnormal area according to the abnormal temperature fluctuation characteristics, effectively locate and extract abnormal temperature change sub-areas, help to quickly lock the problem area and carry out targeted processing, trace the abnormal area to help accurately locate the source of temperature anomaly, and provide guidance and support for subsequent abnormal processing and improvement, perform abnormal cargo node analysis on the abnormal temperature change sub-area, obtain abnormal temperature change cargo nodes, help determine the specific cargo nodes affected, and realize accurate management and monitoring of problem goods, abnormal cargo node analysis helps to quickly identify and process abnormal conditions, and improve the response speed and accuracy of the cold chain cargo state monitoring system.

[0017] Preferably, the specific steps of step S5 are: Step S51: extracting the abnormal cargo boundary image frame based on the abnormal temperature change cargo node positioning; Step S52: performing surface texture detail analysis on the abnormal cargo boundary image frame to generate surface texture details of the abnormal temperature change cargo; Step S53: performing packaging damage detection on the surface texture details of the abnormal temperature change goods to generate abnormal temperature change goods damage detection results.

[0018] The present invention extracts the abnormal cargo boundary image frame based on the abnormal temperature change cargo node positioning, accurately determines the position and range of the abnormal cargo, and provides key information for subsequent analysis and processing. The extraction of the abnormal cargo boundary image frame is helpful to realize the individual positioning and analysis of the abnormal cargo, and provides a data basis for the subsequent surface texture detail analysis. The surface texture detail analysis is performed on the abnormal cargo boundary image frame to generate the surface texture details of the abnormal temperature change cargo, and more comprehensively describe the surface characteristics of the abnormal cargo. The surface texture detail analysis is helpful to deeply understand the surface conditions of the abnormal cargo, and provides an important reference basis for subsequent packaging damage detection and analysis. The packaging damage detection is performed on the surface texture details of the abnormal temperature change cargo, and the abnormal temperature change cargo damage detection results are generated, and the damage of the cargo packaging is discovered in time. The packaging damage detection is helpful to discover the problems of cargo packaging in advance, ensure the integrity and safety of the cargo in transportation, and provide an important basis for the processing and management of abnormal cargo.

[0019] Preferably, the specific steps of step S6 are: Step S61: when the abnormal temperature change cargo damage detection result is that the packaging is not damaged, the cargo temperature state abnormality data is generated; Step S62: Visualizing abnormal temperature of the dynamic temperature distribution twin model based on the abnormal temperature status data of the cargo, thereby generating an abnormal temperature visualization model; Step S63: making abnormal temperature emergency decisions based on the abnormal temperature visualization model and constructing abnormal temperature emergency strategies; Step S64: when the abnormal temperature change cargo damage detection result is packaging damage, a cargo packaging damage signal is generated; Step S65: Perform dynamic transportation path planning based on the cargo packaging damage signal to generate a packaging damage planning path; Step S66: Coordinated optimization of the package damage planning path and abnormal temperature emergency response strategy to build an intelligent cargo status decision model.

[0020] In the present invention, when the abnormal temperature change cargo damage detection result is that the packaging is not damaged, abnormal cargo temperature status data is generated, which is helpful to identify the abnormal temperature of the cargo and take countermeasures in time. The generated abnormal cargo temperature status data provides data support for subsequent abnormal temperature visualization and decision-making, improves the perception and processing efficiency of the abnormal cargo status, and visualizes the abnormal temperature of the dynamic temperature distribution twin model based on the abnormal cargo temperature status data to generate an abnormal temperature visualization model, which is helpful to intuitively display the abnormal temperature distribution. The abnormal temperature visualization model helps users to understand the abnormal temperature distribution more intuitively, and provides visualization support for abnormal temperature emergency decision-making. Abnormal temperature emergency decision-making is made according to the abnormal temperature visualization model, and an abnormal temperature emergency strategy is constructed, which can quickly make emergency treatment when abnormal temperature is found. The construction of abnormal temperature emergency strategy is helpful to deal with potential temperature problems in advance, ensure the quality and safety of goods, and improve the processing efficiency of abnormal situations. When the damage detection result is damaged packaging, a cargo packaging damage signal is generated, which helps to quickly discover the problem of cargo packaging and take timely measures to avoid further damage to the cargo. The generated cargo packaging damage signal provides important information for subsequent dynamic transportation path planning to ensure the safety of the cargo during transportation. Dynamic transportation path planning is carried out based on the cargo packaging damage signal, and a packaging damage planning path is generated, which can avoid potential dangerous areas and ensure the integrity of the cargo during transportation. The formulation of the packaging damage planning path helps to optimize the cargo transportation process, reduce the risk of cargo damage, and improve the safety and efficiency of cargo transportation. The packaging damage planning path and abnormal temperature emergency strategy are coordinated and optimized, and an intelligent cargo status decision model is constructed. It can comprehensively consider temperature anomalies and packaging damage to achieve intelligent decision support. The establishment of an intelligent cargo status decision model helps to improve the overall decision-making ability of the cargo status monitoring system and provide a scientific basis for emergency handling and decision-making during cargo transportation.

[0021] In this specification, a multimodal AIGC cold chain cargo status monitoring system is provided, which is used to execute the multimodal AIGC cold chain cargo status monitoring method as described above, including: The temperature distribution fitting module is used to obtain real-time infrared monitoring images of cold chain goods; perform target visual recognition on the real-time infrared monitoring images of cold chain goods, and fit the temperature distribution of goods in the area to construct a temperature distribution map of goods in the area; The twin model module is used to obtain all-round conventional monitoring images in the cold chain cargo transportation warehouse; perform a three-dimensional layout distribution analysis in the warehouse on the all-round conventional monitoring images according to the cargo temperature distribution map in the area, and perform dynamic temperature distribution rendering to obtain a dynamic temperature distribution twin model; The temperature change trend prediction module is used to calculate the temperature change fluctuation amplitude and predict the temperature change trend at multiple time points for the dynamic temperature distribution twin model, thereby generating a multi-time point temperature change trend prediction map for each sub-area; The abnormal temperature tracing module is used to mine abnormal temperature fluctuations in the multi-time point temperature change trend prediction map of each sub-area, and trace the abnormal area to obtain the abnormal temperature change cargo node; The damage detection module is used to perform surface texture detail analysis based on the abnormal temperature change cargo node, and then perform packaging damage detection to generate abnormal temperature change cargo damage detection results; The cargo status decision module is used to make abnormal temperature emergency decisions and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results, and to build an intelligent cargo status decision model.

[0022] The present invention can realize real-time monitoring and analysis of the temperature distribution of goods by acquiring real-time infrared monitoring images of cold chain goods and performing target visual recognition. The temperature distribution of goods in the area is fitted to construct a temperature distribution map of the goods in the area, which provides basic data support for subsequent temperature distribution analysis and anomaly detection. Based on the temperature distribution map of the goods in the area, a three-dimensional layout distribution analysis in the warehouse is performed to generate a dynamic temperature distribution twin model, which is helpful to fully understand the temperature distribution of the goods in the transport warehouse. The dynamic temperature distribution twin model can provide an important basis for subsequent temperature situation prediction and anomaly detection, and improve the monitoring and early warning capabilities of the goods status. The dynamic temperature distribution twin model is used to calculate the temperature change fluctuation amplitude and predict the temperature change situation at multiple time points to generate a multi-time point temperature change situation prediction map, which provides the ability to predict the temperature change of goods. The multi-time point temperature change situation prediction map helps to monitor the temperature change trend of the goods in real time, and provides more accurate data for the prevention and treatment of abnormal temperature. According to the support, abnormal temperature fluctuations are mined on the multi-time point temperature change trend prediction map to generate abnormal temperature fluctuation characteristics, which is helpful to timely discover abnormal temperature conditions. The abnormal area traceability module can locate the abnormal temperature change cargo nodes, help to quickly deal with abnormal situations, and ensure the quality and safety of the goods. Surface texture detail analysis and packaging damage detection are performed based on the abnormal temperature change cargo nodes to generate abnormal temperature change cargo damage detection results, which is helpful to comprehensively evaluate the status of the goods. The damage of the goods is discovered in time through the damage detection module to improve the timeliness and accuracy of cargo handling. According to the abnormal temperature change cargo damage detection results, abnormal temperature emergency decisions and dynamic transportation route planning are made, and an intelligent cargo status decision-making model is constructed to improve the comprehensive management and decision-making capabilities of the cargo status. The intelligent cargo status decision-making model integrates multimodal data and decision-making processes to provide a scientific basis for emergency handling and decision-making in the cargo transportation process, and improves transportation efficiency and cargo transportation quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1A schematic diagram of the steps of a multi-modal AIGC cold chain cargo status monitoring method of the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0024] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0025] The present application example provides a multimodal AIGC cold chain cargo status monitoring method and system. The execution subject of the multimodal AIGC cold chain cargo status monitoring method and system includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio image management system, an information management system, and a cloud data management system.

[0026] See also Figures 1 to 4 The present invention provides a multi-modal AIGC cold chain cargo status monitoring method, the multi-modal AIGC cold chain cargo status monitoring method comprising the following steps: Step S1: Acquire a real-time infrared monitoring image of cold chain goods; perform target visual recognition on the real-time infrared monitoring image of cold chain goods, perform temperature distribution fitting on the goods in the area, and construct a temperature distribution map of the goods in the area; Step S2: Obtain a full-range conventional monitoring image in the cold chain cargo transport warehouse; perform a three-dimensional layout distribution analysis in the warehouse on the full-range conventional monitoring image according to the cargo temperature distribution map in the area, and perform dynamic temperature distribution rendering to obtain a dynamic temperature distribution twin model; Step S3: Calculate the temperature fluctuation amplitude and predict the temperature change trend at multiple time points for the dynamic temperature distribution twin model, thereby generating a multi-time point temperature change trend prediction map for each sub-region; Step S4: mining abnormal temperature fluctuations on the multi-time point temperature change trend prediction map of each sub-region, and tracing the abnormal region to obtain the abnormal temperature change cargo node; Step S5: Analyze the surface texture details based on the abnormal temperature change cargo node, and then perform packaging damage detection to generate abnormal temperature change cargo damage detection results; Step S6: Make abnormal temperature emergency decisions and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results, and build an intelligent cargo status decision model.

[0027] The present invention obtains real-time infrared monitoring images of cold chain goods and performs target visual recognition to accurately identify goods and obtain real-time temperature data. Constructing a temperature distribution map of goods in the area helps to understand the temperature distribution of goods and discover temperature anomalies or hot spots in advance, thereby ensuring the safety of cold chain transportation of goods. Dynamic temperature distribution rendering is combined with three-dimensional layout distribution analysis in the warehouse to generate a dynamic temperature distribution twin model, providing all-round temperature monitoring and analysis. Through the dynamic temperature distribution twin model, real-time monitoring and prediction of temperature changes of goods in the warehouse are achieved, which helps to ensure the quality of goods. Multi-time point temperature change trend prediction helps to predict the changing trend of cargo temperature and make timely adjustments and response measures. Generating a multi-time point temperature change trend prediction map for each sub-area helps to discover temperature anomalies or hot spots in advance. Abnormal temperature conditions are detected to ensure that the goods are within a safe temperature range. Abnormal temperature fluctuation mining and abnormal area tracing help to accurately locate the cause and scope of abnormal temperature conditions. By generating abnormal temperature change cargo nodes, timely measures are taken to deal with abnormal conditions and ensure the quality of goods. The combination of surface texture detail analysis and package damage detection helps to identify whether the goods are damaged due to abnormal temperature. The abnormal temperature change cargo damage detection results are generated to provide a comprehensive assessment of the cargo status and provide a basis for subsequent processing. Abnormal temperature emergency decision-making and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results help to deal with abnormal conditions in a timely manner and ensure the safety of goods. The construction of an intelligent cargo status decision-making model improves decision-making efficiency and accuracy, and provides intelligent management and guarantee for cold chain cargo transportation.

[0028] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a multi-modal AIGC cold chain cargo status monitoring method of the present invention. In this example, the steps of the method include: Step S1: Acquire a real-time infrared monitoring image of cold chain goods; perform target visual recognition on the real-time infrared monitoring image of cold chain goods, perform temperature distribution fitting on the goods in the area, and construct a temperature distribution map of the goods in the area; In this embodiment, a suitable infrared camera is selected to ensure that it has high resolution and sufficient field of view to fully cover the cold chain cargo area. An infrared camera is installed in the cold chain transport warehouse to ensure that the camera position can effectively capture the temperature information of the cargo. The camera angle is adjusted to avoid reflection and occlusion. The parameters of the infrared camera, such as frame rate, sensitivity and temperature range, are configured to ensure that the temperature changes of the cargo can be accurately captured. An automatic acquisition program is set to regularly acquire real-time infrared monitoring images, and the image data is transmitted to the central processing unit for subsequent analysis. The acquired infrared images are denoised by using methods such as Gaussian filtering or median filtering to reduce noise interference in the image. Contrast enhancement and histogram equalization techniques are applied to improve the clarity and contrast of the infrared image to facilitate subsequent target recognition. A suitable target recognition algorithm is selected, such as YOLO (You Only Look Once) or Faster R-CNN, select the appropriate model for cargo identification according to specific needs, use the selected target recognition algorithm to analyze the processed infrared image, identify various types of cargo in the area, record the bounding box, category and temperature information of each target, verify the accuracy of the recognition result, and ensure that the identified cargo is consistent with the actual situation. If the recognition accuracy is not high, consider adjusting the algorithm parameters or using a more complex model, extract the temperature value of each cargo from the recognition result, and record the location coordinates and temperature information of each target, select a suitable temperature distribution fitting method, such as interpolation (such as Kriging interpolation or bilinear interpolation) or regression analysis method, and construct a temperature distribution map of the cargo in the area based on the extracted temperature data. Use the selected algorithm to spatially fit the temperature to generate a temperature distribution model, visualize the fitted temperature distribution data, and use heat maps or three-dimensional graphics to display the temperature distribution in the area for easy analysis and monitoring.

[0029] Step S2: Obtain a full-range conventional monitoring image in the cold chain cargo transport warehouse; perform a three-dimensional layout distribution analysis in the warehouse on the full-range conventional monitoring image according to the cargo temperature distribution map in the area, and perform dynamic temperature distribution rendering to obtain a dynamic temperature distribution twin model; In this embodiment, a high-resolution surveillance camera is selected to ensure good low-light performance to adapt to the lighting conditions in the warehouse. The installation position and angle of the surveillance camera are designed according to the layout of the cold chain goods to ensure that every corner of the entire transportation warehouse is covered to avoid blind spots. The monitoring system is configured to set a timed shooting or continuous recording mode to ensure that a full range of conventional surveillance images can be obtained in real time. The acquired surveillance images are stored in a database or cloud storage in real time to ensure data security and accessibility. The acquired conventional surveillance images are preprocessed, including denoising, contrast enhancement and color correction to improve image quality. Appropriate image processing and analysis tools are selected, such as OpenCV, MATLAB or deep learning frameworks (such as TensorFlow, PyTorch), and computer vision technology is used to perform target detection on the surveillance images to identify the location and distribution of various types of goods in the warehouse. Based on the recognition results, a three-dimensional spatial layout model of the goods in the warehouse is constructed, the relative position and layout relationship of each goods are recorded, the temperature data of each goods is extracted from the previously generated temperature distribution map of the goods in the area, and it is combined with the spatial layout data of the goods in the warehouse. Suitable three-dimensional rendering tools are selected, such as Blender, Unity or a custom rendering engine, to realize the visualization of dynamic temperature distribution. Based on the spatial layout and temperature data of the goods, a dynamic temperature distribution twin model is constructed, and different temperature areas are represented by heat maps, color mapping, etc. The time dimension is introduced to simulate the dynamic effect of temperature changes over time, and the real-time changes and distribution of temperature are displayed. The generated dynamic temperature distribution twin model is saved to a database or a specific file for subsequent analysis and use, and the dynamic temperature distribution twin model is displayed on a visualization platform, providing an interactive operation interface for users to view the temperature distribution in different time periods.

[0030] Step S3: Calculate the temperature fluctuation amplitude and predict the temperature change trend at multiple time points for the dynamic temperature distribution twin model, thereby generating a multi-time point temperature change trend prediction map for each sub-region; In this embodiment, the temperature data of each sub-area is extracted from the dynamic temperature distribution twin model, including real-time temperature records and historical temperature data, to ensure the integrity and accuracy of the data, handle missing values ​​and outliers, select a suitable fluctuation amplitude calculation method, such as standard deviation, root mean square (RMS), absolute fluctuation amplitude or coefficient of variation (CV), analyze the temperature data of each sub-area, calculate the temperature change fluctuation amplitude, use the selected method to calculate the fluctuation amplitude of each sub-area, and record the results, store the calculated temperature change fluctuation amplitude data of each sub-area in a database for subsequent analysis, select a suitable multi-time point temperature state prediction model, such as ARIMA (autoregressive integrated moving average), LSTM (long short-term memory network) or support vector regression (SVR), divide the collected temperature data into training set and test set, usually using 80% The trained model is used to predict the temperature change trend at multiple time points in each sub-region, and the temperature change data at multiple time points in the future are generated. The prediction results of each sub-region are recorded, including the predicted value, confidence interval and error evaluation. The generated multi-point temperature change trend prediction map of each sub-region and its related data are stored in the database for subsequent analysis and visualization. Visualization tools (such as Matplotlib, Tableau or D3.js) are used to visualize the multi-point temperature change trend prediction map of each sub-region to show the future temperature change trend. An interactive interface is provided to allow users to view and analyze as needed.

[0031] Step S4: mining abnormal temperature fluctuations on the multi-time point temperature change trend prediction map of each sub-region, and tracing the abnormal region to obtain the abnormal temperature change cargo node; In this embodiment, a multi-time point temperature change trend forecast map of each sub-region is extracted from the database to ensure that the data is complete and the format is consistent. A suitable anomaly detection algorithm is selected according to specific needs, such as: Z-score method: based on statistical principles, the Z-score of each data point is calculated to identify abnormal points exceeding the threshold, IQR (interquartile range) method: the quartile of the data is calculated to identify abnormal points exceeding 1.5 times the IQR, machine learning methods: such as Isolation Forest, LOF (local outlier factor), etc., the selected anomaly detection algorithm is applied to the temperature change data of each sub-region to identify abnormal temperature fluctuation points, record the time, amplitude and location of the abnormal fluctuation, store the identified abnormal temperature fluctuation features (such as time, amplitude, sub-region ID) in the database for subsequent analysis, integrate the abnormal temperature fluctuation features with the location information, historical temperature records and packaging status of cold chain goods to form a comprehensive data set, and select a suitable abnormal area traceability analysis method, such as: causal inference: analyze the potential factors leading to the anomaly, symbolic logic reasoning: rule-based reasoning method to find the source of the anomaly, machine learning method: use cluster analysis to identify abnormal areas, and The temperature fluctuation characteristics are combined with the distribution and historical temperature data of the goods to analyze the source of abnormal fluctuations, determine the cargo nodes that affect the abnormal fluctuations, identify the abnormal temperature change cargo nodes, and record the abnormal temperature change cargo node information obtained by tracing (such as cargo ID, location, abnormal fluctuation characteristics) in the database for subsequent processing and management. Visual tools (such as Matplotlib, Tableau or D3.js) are used to display the abnormal temperature fluctuation characteristics and the traceability results of the abnormal area to help relevant personnel quickly understand and respond. An interactive interface is provided to allow users to view the abnormal conditions in different sub-areas. According to the mining and traceability results, a detailed report is generated, including a description of the abnormal situation, the conclusion of the traceability analysis and the recommended countermeasures.

[0032] Step S5: Analyze the surface texture details based on the abnormal temperature change cargo node, and then perform packaging damage detection to generate abnormal temperature change cargo damage detection results; In this embodiment, the extracted image is preprocessed, including denoising, contrast enhancement and color correction, to improve the accuracy of subsequent analysis, and image processing technology (such as local binary pattern LBP, gray level co-occurrence matrix GLCM) is applied to extract surface texture features to obtain detailed information on the surface of the goods, and a suitable feature analysis method is selected, such as principal component analysis (PCA) or support vector machine (SVM), so as to classify and analyze the extracted texture features. The extracted texture features are analyzed to identify abnormal conditions, such as scratches, dents or other surface defects, and a suitable packaging damage detection algorithm is selected, such as: defect detection technology based on image processing (such as edge detection, template matching), deep learning models (such as YOLO, Faster R-CNN) is used for target detection and damage identification. It uses the selected detection algorithm to analyze images of abnormally temperature-changing goods, identify damaged areas on the packaging, including cracks, holes, or tears, etc., record the detected damage location, damage type, and severity, and generate a corresponding test report. It manually reviews the results of automatic detection to ensure the accuracy of detection and reduce false positives and negatives. It stores the package damage detection results (such as damage location, type, and detection image) in the database for subsequent query and management. Based on the results of texture detail analysis and package damage detection, it generates a detailed test report containing test conclusions, recommended treatment measures, and subsequent tracking plans. It uses visualization tools (such as Matplotlib, OpenCV, or dedicated detection software) to display the test results, including annotations of damaged images and heat maps of detection areas, to help relevant personnel quickly understand and respond. It provides a user interface to allow relevant personnel to view and analyze detection data for further decision-making and processing.

[0033] Step S6: Make abnormal temperature emergency decisions and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results, and build an intelligent cargo status decision model.

[0034] In this embodiment, damage detection results of abnormal temperature-changing goods are extracted from the database, including damage type, location, severity and related temperature data, and data related to the transportation environment are collected, such as external temperature, humidity, transportation time and transportation tools. A suitable decision model is selected, such as a rule-based decision tree, a Markov decision process (MDP) or an optimization-based model. According to the damage situation and temperature data, key parameters of emergency decision-making, such as temperature threshold, processing priority and response time, are defined. According to the decision model and parameters, corresponding emergency response measures are formulated, including: reallocating the mode of transportation (such as switching to refrigerated transportation), setting priority categories of goods, selecting appropriate processing locations (such as returning to the warehouse, maintenance center), recording the results of the emergency decision, including decision content, implementation time and responsible person, etc., for subsequent tracking, obtaining real-time data such as the current transportation route, traffic conditions, warehouse location and cargo destination, ensuring the accuracy of information, and selecting a suitable dynamic path planning algorithm. , such as Dijkstra algorithm, A* algorithm or algorithm based on minimum cost flow, considering real-time traffic and transportation conditions, based on the damage detection results and emergency decisions of abnormal temperature-changing goods, re-plan the current transportation route to ensure that the goods arrive at the processing location quickly and safely, generate new transportation routes and estimated arrival times, and record route change information, select suitable machine learning or optimization frameworks, such as TensorFlow, Scikit-learn or custom optimization tools, to build models, integrate historical transportation data, real-time monitoring data and anomaly detection results into decision models to enhance the model's predictive ability, use integrated data sets to train intelligent decision models, adjust model parameters to improve the accuracy of predictions and decisions, use cross-validation and performance evaluation techniques to ensure the effectiveness and stability of the model, test the trained model, evaluate its performance in actual scenarios, and ensure that the model can accurately predict the status of goods and make corresponding decisions.

[0035] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: Perform infrared imaging monitoring on the cold chain cargo transport warehouse to obtain real-time infrared monitoring images of the cold chain cargo; Step S12: Performing target visual recognition on the real-time infrared monitoring image of cold chain goods and marking each cargo node; Step S13: accurately locate the boundary of each cargo node in the region to generate the regional position coordinates of each cargo node; Step S14: performing temperature gradient distribution detection on the real-time infrared monitoring image of cold chain goods, thereby obtaining image temperature gradient distribution characteristics; Step S15: fitting the temperature distribution of cargo in the region to the regional position coordinates of each cargo node according to the temperature gradient distribution characteristics in the region, and constructing a temperature distribution map of cargo in the region.

[0036] In this embodiment, a high-resolution infrared imager is configured to ensure that it can work stably in a cold chain transportation environment. The parameters of the imager, such as focal length, exposure time, and gain, are adjusted to adapt to the lighting conditions and temperature range of the cold chain warehouse. An automatic acquisition system is set to regularly acquire infrared monitoring images of cold chain goods to ensure smooth data transmission. The real-time images are transmitted to a data processing center, and the acquired infrared monitoring images are stored in a database with timestamps and related metadata for subsequent analysis. The real-time infrared monitoring images are preprocessed, including denoising, contrast enhancement, and image normalization, to improve recognition accuracy. A suitable target recognition algorithm is selected, such as a convolutional neural network (CNN), YOLO (You Only LookOnce), or SSD (Single Shot MultiBox Detector), train the target detection model to identify different types of cargo nodes, use the trained model to analyze the infrared monitoring images, identify and mark each cargo node, record the recognition results, including the category and location of each node, select the boundary positioning algorithm, such as edge detection (Canny, Sobel) or contour detection (Contours), to obtain the precise boundary of each cargo node, perform boundary detection on each marked cargo node, generate the regional boundary of each node, calculate the regional location coordinates of each cargo node, ensure the accuracy of the coordinates, record the regional location coordinates of each cargo node in the database for subsequent temperature distribution analysis, extract temperature information from the infrared monitoring images, generate a temperature matrix to represent the temperature value of each pixel, align the temperature value with the image coordinate system to ensure the consistency of the data, and calculate the temperature gradient , use methods such as gradient operators (such as Sobel operator) to obtain the direction and amplitude of temperature change, generate a temperature gradient distribution map, display the temperature change in the area, extract the temperature gradient distribution characteristics, including the highest temperature, the lowest temperature, the average temperature and the temperature change rate, to provide basic data for subsequent analysis, select a suitable temperature distribution fitting model, such as polynomial regression, interpolation or Gaussian process regression, determine the parameters of the fitting model to adapt to the temperature change in the area, fit the temperature distribution of each cargo node according to the obtained regional location coordinates and temperature gradient distribution characteristics, calculate the temperature distribution of each node, generate a temperature distribution map, use visualization tools (such as Matplotlib, Tableau) to display the temperature distribution map of the cargo in the area, clearly identify the temperature change area, compare the fitting results with the actual monitoring data, and verify the accuracy of the fitting.

[0037] In this embodiment, refer to Figure 3, is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Step S21: obtaining a omnidirectional conventional monitoring image in the cold chain cargo transport warehouse; reconstructing the global details of the omnidirectional conventional monitoring image to construct a detailed reconstructed monitoring image; Step S22: segmenting the boundary of each cargo image one by one on the detailed reconstructed monitoring image according to the regional position coordinates of each cargo node, thereby obtaining a plurality of cargo boundary image frames; Step S23: performing a three-dimensional morphological structure analysis on the boundary image frames of the plurality of goods to generate a three-dimensional morphological structure feature of each of the goods; Step S24: performing a three-dimensional layout distribution analysis on the boundary image frames of the multiple goods in the warehouse to obtain three-dimensional layout distribution data of the goods; Step S25: Performing three-dimensional point cloud modeling of the transport warehouse for the three-dimensional morphological structure characteristics of each cargo according to the three-dimensional layout distribution data of the cargo, and constructing a three-dimensional cargo distribution model; Step S26: Render the three-dimensional cargo distribution model with dynamic temperature distribution according to the cargo temperature distribution map in the area to obtain a dynamic temperature distribution twin model.

[0038] In this embodiment, omnidirectional surveillance cameras are installed to ensure coverage of all areas of the cold chain cargo transportation warehouse, obtain clear conventional surveillance images, configure camera parameters such as resolution, frame rate and viewing angle to obtain high-quality images, set automatic acquisition strategies, obtain surveillance images at regular intervals, and store the images in a data processing system to ensure that the acquired images have good lighting conditions to avoid shadows and reflections. Image processing techniques (such as super-resolution reconstruction and image denoising) are used to reconstruct the global details of the acquired surveillance images. Appropriate algorithms such as convolutional neural networks (CNNs) or traditional image enhancement techniques are selected to improve the clarity and details of the images. The surveillance images with reconstructed details are stored in a database for subsequent processing and analysis. The location information of each cargo is extracted from the previous cargo node area location coordinates as the basis for boundary segmentation. Appropriate image segmentation algorithms are selected, such as region-based segmentation (such as GrabCut), edge detection (such as Canny edge detection) or deep learning methods (such as U-Net). The surveillance images with reconstructed details are subjected to global detail reconstruction. Segment the boundaries of each cargo one by one, accurately extract the boundary image frame of each cargo, ensure the accuracy of the segmentation results, record the position and size of each boundary frame, store the boundary image frame of each cargo in the database for subsequent analysis and processing, select appropriate 3D analysis methods, such as volume calculation, shape descriptors (such as Hu moments), principal component analysis, etc., perform 3D morphological structure analysis on the boundary image frames of multiple cargoes, extract the 3D features of each cargo, such as volume, surface area and shape complexity, generate 3D morphological structure feature data for each cargo, record it in the database, verify the extracted 3D features to ensure that they are consistent with the actual cargo shape, select appropriate layout analysis methods, such as cluster analysis, spatial analysis or heat map analysis, perform 3D layout analysis on the boundary image frames of multiple cargoes, identify the distribution of cargoes in space, generate 3D layout distribution data of cargoes, including information such as the distance and relative position between cargoes, store the 3D layout distribution data in the database for subsequent modeling and analysis, select appropriate point cloud modeling tools, such as PCL (Point Cloud Library) or MeshLab, generate a 3D point cloud model of each cargo according to the 3D morphological structure characteristics and three-dimensional layout distribution data of the cargo, merge the 3D point clouds of each cargo, build a 3D cargo distribution model of the entire transport warehouse, verify the accuracy of the generated 3D model, ensure that it is consistent with the actual cargo layout, optimize the model as needed to improve its visualization effect, integrate the cargo temperature distribution map with the 3D cargo distribution model to obtain the temperature information of each cargo in 3D space, and select a suitable rendering technology, such as OpenGL, Unity or Three.js to realize the visualization of dynamic temperature distribution. According to the temperature distribution data of the goods in the area, the dynamic temperature distribution of the three-dimensional goods distribution model is rendered to generate a dynamic temperature distribution twin model, which reflects the temperature changes of the goods in real time. The dynamic temperature distribution twin model is displayed on the visualization platform so that monitoring personnel can view and analyze the status of the goods in real time.

[0039] In this embodiment, the specific steps of step S21 are: Perform in-warehouse ray tracing processing on all-round conventional monitoring images to obtain the ray tracing trajectory of the transport warehouse; Performing light distribution identification on the light tracing trajectory of the transport warehouse to generate light distribution data of the transport warehouse; Perform local brightness detection based on the light distribution data of the transport warehouse and extract the local dark areas in the image; Adaptively enhance the brightness of the local dark areas in the image to construct an adaptive brightness enhanced monitoring image; Perform computer vision detection of cargo contours on adaptive brightness enhanced surveillance images to extract the contours of each cargo; Perform edge detail sharpening on the contour line of each product to obtain the detailed sharpened contour features of each product; Based on the detailed sharpened contour features of each cargo, the global details of the adaptive brightness enhanced surveillance image are reconstructed to construct a detailed reconstructed surveillance image.

[0040] In this embodiment, a conventional monitoring image of a cold chain transport warehouse is obtained to ensure the clarity and integrity of the image. A suitable ray tracing algorithm is selected, such as a ray-based tracing technology. Factors including the position of the light source, light reflection and refraction are considered. Ray tracing processing is performed on the monitoring image to calculate the propagation trajectory of the light in the transport warehouse, generate ray tracing trajectory data, record the path, intensity and reflection characteristics of the light, and select a light distribution recognition method as a result, such as histogram analysis, cluster analysis or heat map generation. According to the ray tracing trajectory data, the distribution of light in the warehouse is analyzed to generate light distribution data, record the intensity, hot spots and shadow areas of the light distribution, and store the light distribution data for subsequent use and analysis. A suitable brightness detection algorithm is selected, such as adaptive threshold processing or regional contrast analysis, to analyze the light distribution data, identify local areas with dim brightness, extract the coordinates and intensity information of these dim areas for subsequent processing, and record the coordinates and features of the local dim areas in the database. An adaptive brightness enhancement algorithm is selected, such as histogram equalization and contrast-limited adaptive histogram equalization (CLAHE), to enhance the local dark areas in the monitoring image to improve their brightness and contrast. degree, ensure that the details of the enhanced image are clear and natural, select a suitable contour detection algorithm, such as Canny edge detection or a segmentation algorithm based on deep learning (such as U-Net), extract the contour of the goods from the surveillance image after adaptive brightness enhancement, identify the contour line of each cargo, record the contour information of each cargo, including contour coordinates and shape features, and store the extracted contour line information in the database for subsequent analysis and processing. Select a suitable edge detail sharpening algorithm, such as Laplace sharpening or high-pass filtering, perform detail sharpening on the contour line of each cargo, enhance the clarity and contrast of the contour, ensure that the contour features after sharpening are more obvious, and facilitate subsequent analysis. Save the sharpened contour features for subsequent image reconstruction. Select a suitable detail reconstruction algorithm, such as super-resolution reconstruction, image fusion technology, or a reconstruction method based on deep learning. Based on the detail sharpening contour features of each cargo, perform global detail reconstruction on the adaptive brightness enhanced surveillance image, generate a surveillance image after detail reconstruction, ensure the overall clarity and detail richness of the image, verify the quality of the reconstructed image, ensure that it meets the expected effect, and store the surveillance image after detail reconstruction in the database for subsequent analysis and application.

[0041] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: performing sub-region spatial division on the dynamic temperature distribution twin model, thereby obtaining a plurality of dynamic temperature sub-regions; Step S32: analyzing the time series variation of the temperature distribution of multiple dynamic temperature sub-regions to obtain the time series variation characteristics of the temperature distribution of each sub-region; Step S33: calculating the temperature change fluctuation amplitude of each sub-region temperature distribution time series change characteristic to generate the temperature change fluctuation amplitude characteristic of each sub-region; Step S34: Evolving the temperature change trend based on the temperature change fluctuation amplitude characteristics, and constructing a temperature change trend evolution curve for each sub-region; Step S35: Perform multi-time point temperature change situation prediction on the temperature change trend evolution curve of each sub-region, so as to generate a multi-time point temperature change situation prediction map of each sub-region. In this embodiment, the created dynamic temperature distribution twin model is loaded to ensure that the model data is complete, and a suitable space division method is selected, such as grid division, cluster analysis, or a division method based on regional characteristics. According to the temperature distribution characteristics and the layout in the warehouse, the dynamic temperature distribution model is divided into multiple sub-areas. Each sub-area should have a clear boundary to ensure that the temperature distribution of the entire transport warehouse is covered. The coordinates, boundaries, and related temperature data of each sub-area are recorded to form a database record of the sub-area. The temperature data of each sub-area is collected to ensure that the data is arranged in chronological order to form a time series data set. A suitable time series analysis method is selected, such as time series analysis, trend analysis, or Fourier transform. The temperature distribution of each sub-area is analyzed for time series changes, and change characteristics such as trends, seasonality, and periodicity are extracted. The time series change characteristics of each sub-area are recorded, including the maximum value, minimum value, and average value, etc. The time series change characteristics of the temperature distribution of each sub-area are stored in the database for subsequent processing. A suitable fluctuation amplitude calculation method is selected, such as standard deviation, root mean square (RMS), or absolute fluctuation amplitude. Analyze the temperature time series data of the region, calculate its temperature fluctuation amplitude, generate the fluctuation amplitude characteristic data of each sub-region, record the value and change trend of the fluctuation amplitude, select a suitable trend evolution modeling method, such as linear regression, smooth curve fitting or polynomial regression, and conduct trend evolution analysis on the temperature change of each sub-region based on the temperature fluctuation amplitude characteristics, construct the temperature change trend evolution curve of each sub-region, record the change law of the trend, and visualize the temperature change trend evolution curve to help analyze and understand the trend of temperature change. Select a suitable prediction model, such as ARIMA, LSTM (long short-term memory network) or regression model, use the temperature change trend data of each sub-region to train the prediction model, optimize the model parameters, apply cross-validation technology to ensure the generalization ability and stability of the model, use the trained model to predict the temperature change trend of each sub-region at multiple time points, generate prediction data of future temperature changes, record the multi-time point temperature change trend prediction map of each sub-region, including the prediction value and its confidence interval, and visualize the multi-time point temperature change trend prediction map to help monitor and manage temperature changes.

[0042] In this embodiment, step S4 includes the following steps: Step S41: performing temperature change mutation analysis on the multi-time point temperature change trend prediction map of each sub-region to identify the temperature change mutation time point; Step S42: mining abnormal temperature fluctuations at the temperature change mutation time point, thereby generating abnormal temperature fluctuation features; Step S43: tracing the source of the abnormal region according to the abnormal temperature fluctuation characteristics, thereby extracting the abnormal temperature change sub-region; Step S44: performing abnormal cargo node analysis on the abnormal temperature change sub-area, thereby obtaining an abnormal temperature change cargo node.

[0043] In this embodiment, a multi-time point temperature change trend forecast map is obtained from each sub-region to ensure the integrity and accuracy of the data, and a suitable temperature change mutation analysis method is selected, such as Cohen's d, CUSUM (cumulative sum control chart) or an anomaly detection algorithm based on machine learning. The selected mutation analysis method is applied to the temperature change data of each sub-region to identify the temperature change mutation time point, record the identified mutation time point, including the temperature change amplitude and the specific time of occurrence, store the relevant features of the mutation time point (such as time, temperature change amplitude) in a database for subsequent analysis, select a suitable abnormal temperature fluctuation mining method, such as a statistical method (such as Z-score) or a machine learning-based anomaly detection model (such as IsolationForest, LOF), perform abnormal fluctuation analysis on the identified temperature change mutation time point, extract abnormal temperature fluctuation features, and record the amplitude, duration, and time of the abnormal fluctuation. The abnormal temperature fluctuation characteristics are saved in the database to facilitate subsequent tracing and analysis. Appropriate tracing methods for abnormal areas are selected, such as causal inference, rule-based tracing or machine learning-based pattern recognition. According to the abnormal temperature fluctuation characteristics, the area corresponding to the warming mutation time point is analyzed to identify the abnormal temperature change sub-area. The specific location and impact range of the abnormal area are determined by combining historical data and real-time monitoring data. The extracted information of the abnormal temperature change sub-area (such as location, temperature characteristics, and fluctuation characteristics) is recorded in the database. Appropriate abnormal cargo node analysis methods are selected, such as target detection based on image processing, cluster analysis or deep learning-based models (such as YOLO and Faster R-CNN). For the abnormal temperature change sub-area, the cargo nodes are analyzed to identify the cargo with abnormal temperature. The characteristics of the abnormal cargo nodes, including cargo type, location and temperature status, are extracted. The information of the abnormal temperature change cargo nodes (such as cargo ID, abnormal temperature value, and location coordinates) is stored in the database to facilitate subsequent processing and decision-making.

[0044] In this embodiment, step S5 includes the following steps: Step S51: extracting the abnormal cargo boundary image frame based on the abnormal temperature change cargo node positioning; Step S52: performing surface texture detail analysis on the abnormal cargo boundary image frame to generate surface texture details of the abnormal temperature change cargo; Step S53: performing packaging damage detection on the surface texture details of the abnormal temperature change goods to generate abnormal temperature change goods damage detection results.

[0045] In this embodiment, the information of abnormal temperature-changing cargo nodes, including their position coordinates and related image data, is obtained from the database, and a suitable bounding box extraction algorithm is selected, such as a region-based segmentation method (such as GrabCut) or a deep learning target detection algorithm (such as YOLO, Faster R-CNN), and the image region where the abnormal cargo node is located is processed to extract the bounding image frame of each abnormal cargo, ensure that the bounding box accurately surrounds the abnormal cargo, record the coordinates and size information of each bounding box, and save the extracted abnormal cargo bounding image frame information to the database for subsequent analysis and processing. A suitable surface texture analysis method is selected, such as local binary pattern (LBP), gray level co-occurrence matrix (GLCM) or convolutional neural network (CNN) for feature extraction, and texture analysis is performed on the extracted abnormal cargo bounding image frame to generate surface texture detail features, and calculate texture feature parameters, such as contrast, entropy, and uniformity. Etc., in order to fully understand the surface state of the goods, store the generated surface texture detail features of the abnormal temperature change goods in the database for subsequent damage detection and analysis, select appropriate packaging damage detection methods, such as image processing-based defect detection technology (such as edge detection, template matching) or deep learning models (such as convolutional neural networks), analyze the surface texture details of the abnormal temperature change goods, identify damage, scratches or other defects on the packaging, record the results of damage detection, including the location, type and severity of the damage, and save the packaging damage detection results (such as damage location, damage type) in the database to provide a basis for subsequent processing and decision-making.

[0046] In this embodiment, step S6 includes the following steps: Step S61: when the abnormal temperature change cargo damage detection result is that the packaging is not damaged, the cargo temperature state abnormality data is generated; Step S62: Visualizing abnormal temperature of the dynamic temperature distribution twin model based on the abnormal temperature status data of the cargo, thereby generating an abnormal temperature visualization model; Step S63: making abnormal temperature emergency decisions based on the abnormal temperature visualization model and constructing abnormal temperature emergency strategies; Step S64: when the abnormal temperature change cargo damage detection result is packaging damage, a cargo packaging damage signal is generated; Step S65: Perform dynamic transportation path planning based on the cargo packaging damage signal to generate a packaging damage planning path; Step S66: Coordinated optimization of the package damage planning path and abnormal temperature emergency response strategy to build an intelligent cargo status decision model.

[0047] In this embodiment, the damage detection results of the abnormal temperature change cargo are obtained from the database to determine whether the "packaging is not damaged". If the packaging is not damaged, the cargo temperature status abnormal data is generated, and the abnormal temperature value, occurrence time, cargo ID and other information are recorded. The cargo temperature status abnormal data is stored in the database for subsequent processing and analysis. Suitable visualization tools or software, such as Matplotlib, Tableau or D3.js, are selected to display dynamic temperature distribution data. Data related to the abnormal temperature status of the cargo are extracted from the dynamic temperature distribution twin model to ensure that the data format is suitable for visualization. Based on the abnormal temperature status data of the cargo, an abnormal temperature visualization model is generated to intuitively display the temperature change of the abnormal area. The visualization effect is enhanced by using heat maps, three-dimensional models or dynamic charts. The generated abnormal temperature visualization model is saved for subsequent decision-making and analysis. Suitable decision models are selected, such as rule-based decision trees, Markov decision processes (MDPs) or optimization-based methods. According to the abnormal temperature visualization model, key parameters for emergency decision-making, such as temperature thresholds, response times and processing priorities, are set. Combined with the abnormal temperature status data and visualization information, an abnormal temperature emergency strategy is constructed, including emergency processing steps, Responsible person and time arrangement, etc., document the constructed abnormal temperature emergency strategy to ensure that relevant personnel can quickly understand and implement it, obtain the damage detection results of abnormal temperature change goods from the database, and determine whether it is "package damage". If the package is damaged, generate a cargo package damage signal, record the damage location, damage degree and other information, and store the cargo package damage signal in the database for subsequent processing and decision-making, select a suitable path planning algorithm, such as Dijkstra algorithm, A* algorithm or algorithm based on minimum cost flow, collect real-time data related to cargo transportation, including current path, traffic conditions, warehouse location, etc., based on Based on the signal of damaged cargo packaging, re-plan the transportation route to ensure that the cargo arrives at the appropriate processing location quickly, generate a damaged packaging planning route, record the new transportation route and estimated arrival time, select a suitable optimization model, such as linear programming, integer programming or multi-objective optimization model, integrate the data of damaged packaging planning path and abnormal temperature emergency strategy, ensure the consistency and integrity of information, consider the mutual influence of the two in the intelligent decision-making model, perform collaborative optimization through optimization algorithm, build an intelligent cargo status decision model, verify the effectiveness of the constructed decision-making model, and make necessary adjustments and optimizations to ensure the speed and accuracy of decision-making.

[0048] In this embodiment, a multimodal AIGC cold chain cargo status monitoring system is provided, which is used to execute the multimodal AIGC cold chain cargo status monitoring method as described above, including: The temperature distribution fitting module is used to obtain real-time infrared monitoring images of cold chain goods; perform target visual recognition on the real-time infrared monitoring images of cold chain goods, and fit the temperature distribution of goods in the area to construct a temperature distribution map of goods in the area; The twin model module is used to obtain all-round conventional monitoring images in the cold chain cargo transportation warehouse; perform a three-dimensional layout distribution analysis in the warehouse on the all-round conventional monitoring images according to the cargo temperature distribution map in the area, and perform dynamic temperature distribution rendering to obtain a dynamic temperature distribution twin model; The temperature change trend prediction module is used to calculate the temperature change fluctuation amplitude and predict the temperature change trend at multiple time points for the dynamic temperature distribution twin model, thereby generating a multi-time point temperature change trend prediction map for each sub-area; The abnormal temperature tracing module is used to mine abnormal temperature fluctuations in the multi-time point temperature change trend prediction map of each sub-area, and trace the abnormal area to obtain the abnormal temperature change cargo node; The damage detection module is used to perform surface texture detail analysis based on the abnormal temperature change cargo node, and then perform packaging damage detection to generate abnormal temperature change cargo damage detection results; The cargo status decision module is used to make abnormal temperature emergency decisions and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results, and to build an intelligent cargo status decision model.

[0049] The present invention can realize real-time monitoring and analysis of the temperature distribution of goods by acquiring real-time infrared monitoring images of cold chain goods and performing target visual recognition. The temperature distribution of goods in the area is fitted to construct a temperature distribution map of the goods in the area, which provides basic data support for subsequent temperature distribution analysis and anomaly detection. Based on the temperature distribution map of the goods in the area, a three-dimensional layout distribution analysis in the warehouse is performed to generate a dynamic temperature distribution twin model, which is helpful to fully understand the temperature distribution of the goods in the transport warehouse. The dynamic temperature distribution twin model can provide an important basis for subsequent temperature situation prediction and anomaly detection, and improve the monitoring and early warning capabilities of the goods status. The dynamic temperature distribution twin model is used to calculate the temperature change fluctuation amplitude and predict the temperature change situation at multiple time points to generate a multi-time point temperature change situation prediction map, which provides the ability to predict the temperature change of goods. The multi-time point temperature change situation prediction map helps to monitor the temperature change trend of the goods in real time, and provides more accurate data for the prevention and treatment of abnormal temperature. According to the support, abnormal temperature fluctuations are mined on the multi-time point temperature change trend prediction map to generate abnormal temperature fluctuation characteristics, which is helpful to timely discover abnormal temperature conditions. The abnormal area traceability module can locate the abnormal temperature change cargo nodes, help to quickly deal with abnormal situations, and ensure the quality and safety of the goods. Surface texture detail analysis and packaging damage detection are performed based on the abnormal temperature change cargo nodes to generate abnormal temperature change cargo damage detection results, which is helpful to comprehensively evaluate the status of the goods. The damage of the goods is discovered in time through the damage detection module to improve the timeliness and accuracy of cargo handling. According to the abnormal temperature change cargo damage detection results, abnormal temperature emergency decisions and dynamic transportation route planning are made, and an intelligent cargo status decision-making model is constructed to improve the comprehensive management and decision-making capabilities of the cargo status. The intelligent cargo status decision-making model integrates multimodal data and decision-making processes to provide a scientific basis for emergency handling and decision-making in the cargo transportation process, and improves transportation efficiency and cargo transportation quality.

[0050] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0051] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A multi-modal AIGC cold chain cargo status monitoring method, characterized in that: The following steps are involved: Step S1: Acquire a real-time infrared monitoring image of cold chain goods; perform target visual recognition on the real-time infrared monitoring image of cold chain goods, perform temperature distribution fitting on the goods in the area, and construct a temperature distribution map of the goods in the area; Step S2: Obtain a full-range conventional monitoring image in the cold chain cargo transport warehouse; perform a three-dimensional layout distribution analysis in the warehouse on the full-range conventional monitoring image according to the cargo temperature distribution map in the area, and perform dynamic temperature distribution rendering to obtain a dynamic temperature distribution twin model; Step S3: Calculate the temperature fluctuation amplitude and predict the temperature change trend at multiple time points for the dynamic temperature distribution twin model, thereby generating a multi-time point temperature change trend prediction map for each sub-region; Step S4: mining abnormal temperature fluctuations on the multi-time point temperature change trend prediction map of each sub-region, and tracing the abnormal region to obtain the abnormal temperature change cargo node; Step S5: Surface texture detail analysis is performed based on the abnormal temperature change cargo node, and then packaging damage detection is performed to generate abnormal temperature change cargo damage detection results; Step S6: Make abnormal temperature emergency decisions and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results, and build an intelligent cargo status decision model.

2. The multi-modal AIGC cold chain cargo status monitoring method according to claim 1 is characterized in that: The specific steps of step S1 are: Step S11: Perform infrared imaging monitoring on the cold chain cargo transport warehouse to obtain real-time infrared monitoring images of the cold chain cargo; Step S12: Performing target visual recognition on the real-time infrared monitoring image of cold chain goods and marking each cargo node; Step S13: accurately locate the boundary of each cargo node in the region to generate the regional position coordinates of each cargo node; Step S14: performing temperature gradient distribution detection on the real-time infrared monitoring image of cold chain goods, thereby obtaining image temperature gradient distribution characteristics; Step S15: fitting the temperature distribution of cargo in the region to the regional position coordinates of each cargo node according to the temperature gradient distribution characteristics in the region, and constructing a temperature distribution map of cargo in the region.

3. The method according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: obtaining a omnidirectional conventional monitoring image in the cold chain cargo transport warehouse; reconstructing the global details of the omnidirectional conventional monitoring image to construct a detailed reconstructed monitoring image; Step S22: segmenting the boundary of each cargo image one by one on the detailed reconstructed monitoring image according to the regional position coordinates of each cargo node, thereby obtaining a plurality of cargo boundary image frames; Step S23: performing a three-dimensional morphological structure analysis on the boundary image frames of the plurality of goods to generate a three-dimensional morphological structure feature of each of the goods; Step S24: performing a three-dimensional layout distribution analysis on the boundary image frames of the multiple goods in the warehouse to obtain three-dimensional layout distribution data of the goods; Step S25: Performing three-dimensional point cloud modeling of the transport warehouse for the three-dimensional morphological structure characteristics of each cargo according to the three-dimensional layout distribution data of the cargo, and constructing a three-dimensional cargo distribution model; Step S26: Render the three-dimensional cargo distribution model with dynamic temperature distribution according to the cargo temperature distribution map in the area to obtain a dynamic temperature distribution twin model.

4. The multi-modal AIGC cold chain cargo status monitoring method according to claim 3 is characterized in that: The specific steps of step S21 are: Perform in-warehouse ray tracing processing on all-round conventional monitoring images to obtain the ray tracing trajectory of the transport warehouse; Performing light distribution identification on the light tracing trajectory of the transport warehouse to generate light distribution data of the transport warehouse; Perform local brightness detection based on the light distribution data of the transport warehouse and extract the local dark areas in the image; Adaptively enhance the brightness of the local dark areas in the image to construct an adaptive brightness enhanced monitoring image; Perform computer vision detection of cargo contours on adaptive brightness enhanced surveillance images to extract the contours of each cargo; Perform edge detail sharpening on the contour line of each product to obtain the detailed sharpened contour features of each product; Based on the detailed sharpened contour features of each cargo, the global details of the adaptive brightness enhanced surveillance image are reconstructed to construct a detailed reconstructed surveillance image.

5. The multi-modal AIGC cold chain cargo status monitoring method according to claim 1 is characterized in that: The specific steps of step S3 are: Step S31: performing sub-region spatial division on the dynamic temperature distribution twin model, thereby obtaining a plurality of dynamic temperature sub-regions; Step S32: analyzing the time series variation of the temperature distribution of multiple dynamic temperature sub-regions to obtain the time series variation characteristics of the temperature distribution of each sub-region; Step S33: calculating the temperature change fluctuation amplitude of each sub-region temperature distribution time series change characteristic to generate the temperature change fluctuation amplitude characteristic of each sub-region; Step S34: Evolving the temperature change trend based on the temperature change fluctuation amplitude characteristics, and constructing a temperature change trend evolution curve for each sub-region; Step S35: Perform multi-time point temperature change situation prediction on the temperature change trend evolution curve of each sub-region, so as to generate a multi-time point temperature change situation prediction map of each sub-region.

6. The multi-modal AIGC cold chain cargo status monitoring method according to claim 1 is characterized in that: The specific steps of step S4 are: Step S41: performing temperature change mutation analysis on the multi-time point temperature change trend prediction map of each sub-region to identify the temperature change mutation time point; Step S42: mining abnormal temperature fluctuations at the temperature change mutation time point, thereby generating abnormal temperature fluctuation features; Step S43: tracing the source of the abnormal region according to the abnormal temperature fluctuation characteristics, thereby extracting the abnormal temperature change sub-region; Step S44: performing abnormal cargo node analysis on the abnormal temperature change sub-area, thereby obtaining an abnormal temperature change cargo node.

7. The multi-modal AIGC cold chain cargo status monitoring method according to claim 1 is characterized in that: The specific steps of step S5 are: Step S51: extracting the abnormal cargo boundary image frame based on the abnormal temperature change cargo node positioning; Step S52: performing surface texture detail analysis on the abnormal cargo boundary image frame to generate surface texture details of the abnormal temperature change cargo; Step S53: performing packaging damage detection on the surface texture details of the abnormal temperature change goods to generate abnormal temperature change goods damage detection results.

8. The multi-modal AIGC cold chain cargo status monitoring method according to claim 1 is characterized in that: The specific steps of step S6 are: Step S61: when the abnormal temperature change cargo damage detection result is that the packaging is not damaged, the cargo temperature state abnormality data is generated; Step S62: Visualizing abnormal temperature of the dynamic temperature distribution twin model based on the abnormal temperature status data of the cargo, thereby generating an abnormal temperature visualization model; Step S63: making abnormal temperature emergency decisions based on the abnormal temperature visualization model and constructing abnormal temperature emergency strategies; Step S64: when the abnormal temperature change cargo damage detection result is packaging damage, a cargo packaging damage signal is generated; Step S65: Perform dynamic transportation path planning based on the cargo packaging damage signal to generate a packaging damage planning path; Step S66: Coordinated optimization of the package damage planning path and abnormal temperature emergency response strategy to build an intelligent cargo status decision model.

9. A multi-modal AIGC cold chain cargo status monitoring system, characterized in that: Used to execute the multimodal AIGC cold chain cargo status monitoring method as claimed in claim 1, comprising: The temperature distribution fitting module is used to obtain real-time infrared monitoring images of cold chain goods; perform target visual recognition on the real-time infrared monitoring images of cold chain goods, and fit the temperature distribution of goods in the area to construct a temperature distribution map of goods in the area; The twin model module is used to obtain all-round conventional monitoring images in the cold chain cargo transportation warehouse; perform a three-dimensional layout distribution analysis in the warehouse on the all-round conventional monitoring images according to the cargo temperature distribution map in the area, and perform dynamic temperature distribution rendering to obtain a dynamic temperature distribution twin model; The temperature change trend prediction module is used to calculate the temperature change fluctuation amplitude and predict the temperature change trend at multiple time points for the dynamic temperature distribution twin model, thereby generating a multi-time point temperature change trend prediction map for each sub-area; The abnormal temperature tracing module is used to mine abnormal temperature fluctuations in the multi-time point temperature change trend prediction map of each sub-area, and trace the abnormal area to obtain the abnormal temperature change cargo node; The damage detection module is used to perform surface texture detail analysis based on the abnormal temperature change cargo node, and then perform packaging damage detection to generate abnormal temperature change cargo damage detection results; The cargo status decision module is used to make abnormal temperature emergency decisions and dynamic transportation route planning based on the abnormal temperature change cargo damage detection results, and to build an intelligent cargo status decision model.

Images