A container stacking monitoring system and method

Abstract

This invention relates to a container stacking monitoring system and method, belonging to the field of container management technology. A processing unit constructs a stacking model corresponding to non-standard cargo based on stacking image data and an artificial intelligence model. The system adjusts the movement path of the non-standard cargo based on the stacking scheme generated by the stacking model. A stacking machinery unit executes the stacking task based on the stacking scheme and the movement path of the non-standard cargo sent by the processing unit. During the stacking process, the stacking model adjusts the stacking scheme based on the geometric, vibration, weight, and offset characteristics of the cargo fed back by the stacking machinery unit. Addressing the technical problem of reducing the impact of non-standard cargo stacking on the stability of container stacking, this invention combines the actual characteristic parameters obtained by the stacking machinery unit along the movement path of the non-standard cargo to update and adjust the stacking scheme, thereby achieving a more stable and safer stacking effect.

Description

Technical Field

[0001] This invention relates to the field of container management technology, and in particular to a container stacking monitoring system and method. Background Technology

[0002] In modern terminals, container stacking is increasingly reliant on intelligent routing systems to improve efficiency, reduce delays, and optimize the entire process. These systems collect and analyze terminal operation data in real time, including container size, weight, destination, estimated arrival time, and vessel loading / unloading schedules. This data is used to calculate optimal container stacking locations and transport routes. Based on the collected data, the intelligent routing system dynamically generates optimal container movement paths. This means that paths are adjusted according to real-time conditions (such as current container layout, the location and status of handling equipment) to avoid congestion and delays. The intelligent routing system uses predictive algorithms to forecast future container traffic and handling equipment demand, thereby planning routes and resource allocation in advance, reducing waiting time and increasing throughput. Container movement typically involves equipment such as automated guided vehicles (AGVs), automated stacker cranes (ASCs), and automated guided vehicles (AGVs). These automated devices operate according to the routes and scheduling instructions provided by the intelligent routing system, ensuring that containers are quickly and accurately unloaded from ships, stacked on the terminal, or loaded onto other transport vehicles. Stacking involves not only spatial routing planning but also vertical stacking planning. The intelligent routing system considers container retrieval frequency, prioritizing the stacking of containers about to be used to reduce movement and time. It ensures the continuity and coordination of the entire operation. The system coordinates the timing of different equipment and operations to ensure a smooth transition from one operation to the next, thereby reducing container dwell time at the terminal. Furthermore, the system monitors equipment status in real time; upon detecting malfunctions or anomalies, it can quickly replan routes and adjust the order of operations to minimize impact on the overall operation. Operators monitor the intelligent routing system through a control interface and can intervene manually when necessary, such as modifying routes or adjusting priorities, to address unforeseen changes or emergencies.

[0003] However, current terminal intelligent systems focus on intelligent path planning, neglecting the unique characteristics of cargo when stacking within containers. This results in some cargo being unsuitable for intelligent mechanical stacking and still requiring manual stacking. Some cargo with special shapes or sizes may not be accurately identified and handled by the intelligent planning system. For example, non-standard sized cargo, specially shaped equipment, or machine parts may require manual intervention to be correctly stacked in containers. For fragile or easily damaged cargo, the intelligent planning system may not provide sufficient protection to prevent damage during stacking. Manual stacking allows for more careful handling based on the characteristics and needs of the cargo, ensuring its safety. Some cargo may have special handling requirements, necessitating personalized stacking and arrangement based on specific circumstances. For example, flammable, explosive, or toxic cargo requires special handling and isolation measures, which the intelligent planning system may not be able to meet. Therefore, how to improve current intelligent planning systems to implement differentiated stacking schemes for special cargoes remains a technical problem that current intelligent planning systems have not yet solved.

[0004] Chinese patent application CN117049199A discloses a container loading method, including the following steps: A1. Obtaining loading order information, which includes the container dimensions, the type of boxed goods to be loaded, and the dimensions and quantity of boxed goods for each type; A2. Based on the container dimensions and the dimensions of the boxed goods for each type, obtaining the optimal single-layer stacking configuration and corresponding layer width for each type of boxed goods; A3. Based on the container dimensions, the dimensions and quantity of the boxed goods for each type of box, and the dimensions and quantity of the boxed goods for each type of box, obtaining the optimal single-layer stacking configuration and corresponding layer width for each type of boxed goods. The optimal single-layer stacking configuration for box-shaped goods of various container types is determined by identifying the number of stacks, layers, and individual goods that can be formed for each container type. These are denoted as the number of stacks, layers, and individuals, respectively, and linked lists are created for stacking, layers, and individuals. A4. Based on the principle of prioritizing filling the gaps at the top of the stacks, the packing position of each container is determined according to the stacking list, layer list, individual list, the dimensions of each container type, the optimal single-layer stacking configuration, and the corresponding layer width, resulting in a packing scheme. A5. Packing is performed according to the packing scheme. However, this patent application only addresses stacking schemes for special container bodies and cannot implement corresponding stacking schemes for irregular or fragile goods.

[0005] Therefore, the present invention aims to provide a container stacking monitoring system that can monitor the stacking path and location of special goods, issue timely warnings, and provide construction personnel with reasonable stacking adjustment schemes.

[0006] Furthermore, on the one hand, there are differences in understanding among those skilled in the art; on the other hand, the applicant studied a large number of documents and patents when making this invention, but due to space limitations, not all details and contents were listed in detail. However, this does not mean that the present invention does not possess the features of these prior art. On the contrary, the present invention already possesses all the features of the prior art, and the applicant reserves the right to add relevant prior art to the background art. Summary of the Invention

[0007] Current intelligent terminal systems focus on intelligent path planning, neglecting the unique characteristics of cargo during container stacking. This results in some cargo being unsuitable for intelligent mechanical stacking and still requiring manual stacking. Some cargo with special shapes or sizes may not be accurately identified and handled by the intelligent planning system. For example, non-standard sized cargo, specially shaped equipment, or machine parts may require manual intervention to be correctly stacked in containers. For fragile or easily damaged cargo, the intelligent planning system may not provide sufficient protection to prevent damage during stacking. Manual stacking allows for more careful handling based on the characteristics and needs of the cargo, ensuring its safety. Some cargo may have special handling requirements, necessitating personalized stacking and arrangements based on specific circumstances. For example, flammable, explosive, or toxic cargo requires special handling and isolation measures, which the intelligent planning system may not be able to meet. Therefore, how to improve current intelligent planning systems to implement differentiated stacking schemes for special cargoes remains a technical problem that current intelligent planning systems have not yet solved.

[0008] Existing technologies have already developed technical solutions for adjusting palletizing methods in real time based on incoming material information and stack shape information. For example, Chinese patent application CN117800099A discloses a palletizing method, which includes: acquiring observation information of the current task environment, including incoming material information of the incoming material area and stack shape information of the palletizing area; outputting palletizing decision information, including stacking position and stacking direction, based on the incoming material information and the stack shape information using a trained reinforcement learning model; and controlling a robotic arm to perform a palletizing task based on the stacking position and stacking direction. The trained reinforcement learning model is obtained by training on action information of the palletizing task, observation information before the action is executed, and reward information after the action is executed. This technical solution, through the trained reinforcement learning model, can provide palletizing decision information for non-standardized and disordered goods, enabling automated operation in complex palletizing scenarios with randomly changing incoming box sizes and order, improving palletizing efficiency for non-standardized and disordered goods, reducing reliance on manual handling, and lowering labor costs. This technical solution primarily emphasizes the matching degree between the robotic arm's posture adjustment and the incoming material's posture. This allows for timely adjustments to the robotic arm's gripping actions based on different incoming material information, thereby reducing the number of adjustments required and achieving the minimum robotic arm adjustment parameters needed to meet the requirements of incoming material information, stacking tasks, and stack shape information. However, this technical solution cannot address the impact of non-standard goods, such as those with special shapes or fragile items, on the actual stacking stability. It only pursues the matching degree of the goods' external dimensions, failing to consider factors affecting the overall safety of cargo transportation, such as the instability of the overall stacking model caused by the shift of the goods' center of gravity. Therefore, it cannot meet the actual transportation requirements of non-standard and fragile goods.

[0009] To address the shortcomings of existing technologies, this invention provides a container stacking monitoring system, comprising an image acquisition unit, a processing unit, and a stacking machinery unit. The image acquisition unit acquires stacking image data corresponding to non-standard cargo; the processing unit constructs a stacking model corresponding to the non-standard cargo based on the stacking image data and an artificial intelligence model; the processing unit adjusts the movement path of the non-standard cargo based on the stacking scheme generated by the stacking model; the stacking machinery unit executes the stacking task based on the stacking scheme and the movement path of the non-standard cargo sent by the processing unit; wherein, during the stacking process, the stacking model adjusts the stacking scheme based on the geometric features, vibration features, weight features, and offset features of the cargo fed back by the stacking machinery unit. Unlike existing technologies, the stacking task executed by the stacking machinery unit of this invention is jointly determined by the stacking scheme generated by the stacking model and the movement path of the non-standard cargo, wherein the movement path of the non-standard cargo is further adjusted according to the stacking scheme generated by the stacking model, and during the stacking process, the stacking scheme can be adjusted in real time based on the geometric features, vibration features, weight features, and offset features of the cargo fed back by the stacking machinery unit. Based on the aforementioned distinguishing technical features, the problem this invention aims to solve can include: how to adjust subsequent stacking schemes according to the actual impact of non-standard goods on container stacking during the stacking process, so as to reduce the impact of non-standard goods stacking on container stacking stability, thereby improving the safety of cargo stacking and transportation within containers. Specifically, this invention achieves rapid stacking model construction by inputting real-time collected stacking image data into an artificial intelligence model, thereby achieving synchronous response between stacking image data and the stacking model. This improves the matching degree between the stacking model and the actual cargo stacking state, and also enables the stacking scheme generated based on the stacking model to match the actual cargo stacking requirements. During the stacking process of non-standard goods, the stacking mechanical unit executes the stacking task according to the preset stacking scheme and the movement path of the non-standard goods. At this time, the actual characteristic parameters of the non-standard goods can be obtained through the stacking mechanical unit. These actual characteristic parameters will affect the actual stacking effect of the non-standard goods, and these effects are unpredictable by the initial stacking scheme. If the stacking process is executed according to the initial stacking scheme, it will lead to the instability of the overall stacking structure. This invention updates and adjusts the stacking scheme by combining actual characteristic parameters obtained by the stacking mechanical unit along the movement path of non-standard goods, thereby achieving a more stable and safer stacking effect. This invention can monitor the stacking of non-standard goods and update the stacking scheme in real time to rationally stack special goods and improve stacking efficiency. Specifically, this invention can automatically adjust the stacking layout, optimize the position of goods, reduce movement and collisions, and ensure the safe and stable loading of goods inside containers, thereby improving the efficiency and reliability of the entire transportation process.

[0010] According to a preferred embodiment, before stacking, the stacking model adjusts the stacking scheme based on spatial image data of the container acquired by the image acquisition unit, and determines the stacking influence relationship between the stacked goods and non-standard goods based on the type of the stacked goods and their transport marking information, thereby adjusting the stacking parameters of the non-standard goods. This invention can intelligently adjust the stacking parameters of non-standard goods according to the characteristics and interrelationships of different goods to achieve the optimal stacking layout.

[0011] According to a preferred embodiment, after stacking, the stacking model calculates the stability parameters of the goods based on the stacking image data of the stacking pile collected by the image acquisition unit, and performs similarity calculation between the image data of the stacking pile and image samples of stacking pile collapse cases to determine the safety of the stacking pile.

[0012] The stacking model utilizes image data of the stacked goods acquired by the image acquisition unit to comprehensively assess the stability of the goods. By calculating the stability parameters of the goods and comparing them with image samples from existing stacked goods collapse cases, potential safety hazards can be identified in a timely manner, allowing necessary measures to be taken to ensure the safety and stability of the stacked goods.

[0013] According to a preferred embodiment, the stacking model determines the core position of each stacking state in the stacking scheme. During the stacking process, the stacking model extracts the placement features of the goods at the core position of the stacking acquired by the image acquisition unit, and calculates the stability of the stacking pile based on the placement features of the goods at the core position of the stacking.

[0014] By dynamically adjusting the stacking scheme based on real-time data, this invention can respond promptly to changes in cargo placement and the impact of the external environment, thereby ensuring the smooth progress of the entire stacking process and the final loading safety.

[0015] According to a preferred embodiment, during the stacking process, the stacking model verifies the stability of the non-standard goods at the core stacking location based on their geometric, vibration, weight, and offset characteristics. Furthermore, the stacking scheme is adjusted in real-time based on the stability of the non-standard goods at the core stacking location. By analyzing these characteristics and calculating the stability of the stack, this invention can provide timely feedback and guidance during actual stacking to ensure proper placement of goods and the stability of the stack, thereby improving loading efficiency and safety.

[0016] According to a preferred embodiment, the processing unit further includes a behavior recognition model. After the stacking is completed, the behavior recognition model identifies intervention behaviors during the manual inspection of the stacking pile based on the image data collected by the image acquisition unit. The behavior recognition model sends intervention behavior tags to the stacking model, so that the stacking model associates the image data of the stacking scheme and intervention time with the intervention behavior tags. By monitoring and recognizing various intervention behaviors during the manual inspection process in real time, such as rearranging goods or moving goods, this invention can provide important information for subsequent stacking optimization and can also provide an optimization basis for optimizing the stacking model.

[0017] According to a preferred embodiment, the scrambling model extracts corresponding image data based on the intervention behavior markers sent by the behavior recognition model, and compares the scrambling parameters of the intervention position before and after the intervention based on the extracted image data corresponding to the intervention behavior markers. The scrambling model sends the scrambling parameters of the intervention position before and after the intervention, as well as the image data, to the terminal for manual confirmation. If the terminal provides feedback that the intervention is effective, the scrambling model optimizes the scrambling scheme based on the scrambling parameters of the intervention position after the intervention.

[0018] By sending the placement parameters and image data before and after the intervention to the terminal for manual confirmation, and based on the effective intervention information fed back by the terminal, the placement model can effectively process the intervention and optimize the placement scheme, ensuring that the intervention behavior can be identified and processed in a timely manner, and improving the automation and efficiency of the placement process.

[0019] According to a preferred embodiment, the stacking model updates the core position of the stacking in real time based on the state of the stacking pile of non-standard goods.

[0020] By continuously monitoring and analyzing the stacking status and stability of non-standard goods, the stacking model can dynamically adjust the core position of the stacking to adapt to changes in the placement of goods. This ensures that the system can update and optimize the core position in a timely manner during the stacking process, thereby improving the rationality and stability of the stacking layout.

[0021] The present invention provides a container stacking monitoring method from a second aspect. The method includes: collecting stacking image data corresponding to non-standard goods; constructing a stacking model corresponding to the non-standard goods based on the stacking image data and an artificial intelligence model; adjusting the movement path of the non-standard goods based on the stacking scheme generated by the stacking model; and executing a stacking task based on the stacking scheme and the movement path of the non-standard goods sent by the processing unit. During the stacking process, the stacking scheme is adjusted based on the geometric characteristics, vibration characteristics, weight characteristics, and offset characteristics of the goods.

[0022] By collecting image data corresponding to non-standard cargo, a comprehensive perception of the internal layout of the container is achieved. This allows for accurate acquisition of the position, status, and layout of each item within the container, providing essential data support for the subsequent stacking process. Utilizing stacking image data and artificial intelligence models, a stacking model corresponding to non-standard cargo is constructed. Based on image data and deep learning model training, the type, shape, and position of non-standard cargo can be accurately identified, and a corresponding stacking model can be established, providing a reliable basis for the generation and optimization of subsequent stacking schemes. By adjusting the movement path of non-standard cargo, dynamic optimization of the stacking scheme is achieved. The position and layout of cargo can be flexibly adjusted according to the geometric, vibration, weight, and offset characteristics of different cargoes to maximize the loading efficiency and stability of the container. This invention comprehensively considers multiple cargo characteristics and dynamically optimizes the stacking scheme to ensure stable cargo placement and safe container loading.

[0023] According to a preferred embodiment, the method further includes: adjusting the stacking scheme based on spatial image data of the container, and determining the stacking influence relationship between the stacked goods and non-standard goods based on the type of the stacked goods and their transport marking information, thereby adjusting the stacking parameters of the non-standard goods. This invention intelligently adjusts the stacking scheme according to the actual internal space of the container, improving the efficiency and accuracy of the stacking process. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of an unstacked scenario in the container stacking monitoring system provided by the present invention;

[0025] Figure 2 This is a schematic diagram of a stacked container scene provided by the container stacking monitoring system of the present invention;

[0026] Figure 3 This is a schematic diagram of a scenario requiring manual intervention in the container stacking monitoring system provided by the present invention;

[0027] Figure 4 This is a schematic diagram of the stacking mechanical unit provided by the present invention;

[0028] Figure 5 This is a logical schematic diagram of the container stacking monitoring system provided by the present invention;

[0029] Figure 6 This is a flowchart illustrating the container stacking monitoring system provided by the present invention;

[0030] Figure 7 This is a schematic diagram of the steps of the container stacking monitoring method provided by the present invention.

[0031] List of reference numerals

[0032] 100: Image acquisition unit; 110: Non-standard goods; 120: Stacking stack; 200: Processing unit; 210: Stacking model; 220: Behavior recognition model; 300: Stacking mechanical unit; 310: Fixture; 320: Weight sensor; 330: Vibration sensor; 400: Terminal; 500: Person. Detailed Implementation

[0033] The following is a detailed explanation with reference to the accompanying drawings.

[0034] The image acquisition unit 100 comprises two parts: one part is installed inside or around the container to acquire image data of the goods; the other part is installed on the stacking machinery unit 300 to acquire real-time image data of the stacking status of the stacking stack 120.

[0035] The processing unit 200 is typically located in a control center or computer server room near the container terminal or dock. The processing unit 200 is used to run the stacking model 210 and the behavior recognition model 220. The processing unit 200 is a combination of hardware such as a dedicated integrated chip, server, and processor capable of running the stacking model 210 and the behavior recognition model 220 of this invention.

[0036] The stacking machinery unit 300 is located inside or around the container and is typically an automated device or robotic arm system. For example... Figure 4 As shown, the stacking mechanical unit 300 can also be an intelligent robot for stacking goods. The intelligent robot includes an image acquisition component as part of the image acquisition unit 100, which acquires the geometric and offset features of the goods. The image acquisition component may include sensors such as cameras or lidar to acquire the position, shape, and offset of the goods in real time. A vibration sensor 330 is installed on the gripper 310 of the intelligent robot to monitor the vibration parameters and changes of the goods during transportation and stacking to ensure the stability of the goods. A weight sensor 320 is installed on the gripper 310 of the intelligent robot to monitor the weight of the goods in real time. The intelligent robot also includes a motion control unit, including motors, drivers, encoders, etc., for controlling the robot's movement path and actions. Based on the stacking scheme and path planning generated by the processing unit 200, the motion control unit controls the joint movements and end effector actions of the intelligent robot to achieve precise stacking of goods. The intelligent robot also includes a main control unit, including a main controller, processor, sensor interfaces, etc., for controlling and scheduling the entire intelligent robot system. The main control unit receives the stacking scheme and adjustment instructions sent by the processing unit 200, and uses programming algorithms to coordinate and control the various components of the intelligent robot to complete the stacking task.

[0037] The image acquisition unit 100 transmits the acquired image data to the processing unit 200 via a data cable or network connection. The processing unit 200 sends the generated staging scheme to the staging machine unit 300 via a data cable or network connection. The staging machine unit 300 can receive the staging scheme sent by the processing unit 200 via a wired or wireless connection and execute the corresponding staging task.

[0038] Shipping marking information includes cargo category, name, weight, dimensions, destination marking, fragile item marking, moisture-proof marking, stacking restrictions, loading and unloading instructions, etc. Preferably, the shipping marking information may also include barcode or QR code information on the surface of the packaging box. Example 1

[0039] Current intelligent terminal systems focus on intelligent path planning, neglecting the unique characteristics of cargo during container stacking. This results in some cargo being unsuitable for intelligent mechanical stacking and still requiring manual stacking. Some cargo with special shapes or sizes may not be accurately identified and handled by the intelligent planning system. For example, non-standard sized cargo, specially shaped equipment, or machine parts may require manual intervention to be correctly stacked in containers. For fragile or easily damaged cargo, the intelligent planning system may not provide sufficient protection to prevent damage during stacking. Manual stacking allows for more careful handling based on the characteristics and needs of the cargo, ensuring its safety. Some cargo may have special handling requirements, necessitating personalized stacking and arrangements based on specific circumstances. For example, flammable, explosive, or toxic cargo requires special handling and isolation measures, which the intelligent planning system may not be able to meet. Therefore, how to improve current intelligent planning systems to implement differentiated stacking schemes for special cargoes remains a technical problem that current intelligent planning systems have not yet solved.

[0040] To address the shortcomings of existing technologies, this invention provides a container stacking monitoring system, such as... Figure 1 As shown, it includes an image acquisition unit 100, a processing unit 200, and a stacking machinery unit 300. The image acquisition unit 100 acquires stacking image data corresponding to the non-standard goods 110.

[0041] The processing unit 200 constructs a stacking model 210 corresponding to the non-standard goods 110 based on the stacking image data and artificial intelligence model. The processing unit 200 adjusts the movement path of the non-standard goods 110 based on the stacking scheme generated by the stacking model 210.

[0042] Specifically, a large amount of sample data on non-standard goods 110 is preprocessed, including data cleaning, standardization, and normalization, to facilitate the training and analysis of the stacking model 210. The sample data on non-standard goods 110 includes, but is not limited to, the geometric characteristics (such as size and shape), weight characteristics, vibration characteristics, and offset characteristics of the goods. Data collection can be accomplished using devices such as sensors and cameras.

[0043] After sample data preprocessing, feature extraction is performed to extract useful information and construct features that are helpful to the stacking model 210 from the original sample data. For example, computer vision models are used to extract features such as contour, volume, and centroid position from image data.

[0044] Based on the characteristics of the task, a suitable artificial intelligence (AI) model is selected. This AI model can be a decision tree, random forest, deep learning model, etc. After selecting the AI ​​model, it is trained using the previously processed sample data. The goal of the training is to enable the AI ​​model to predict the optimal stacking scheme based on the characteristics of the goods. For example, a convolutional neural network (CNN) can be trained using a data sample set to ensure it can handle various non-standard goods 110. The model constructed in this way is the stacking model 210.

[0045] The optimized stacking model 210 will be integrated into the processing unit 200, enabling the processing unit 200 to generate a stacking scheme based on the image data of the non-standard goods 110 and adjust the movement path of the non-standard goods 110. Furthermore, the stacking model 210 also needs to be integrated with the stacking machinery unit 300 to allow for real-time adjustment of the stacking scheme based on data fed back from the stacking machinery unit 300.

[0046] After the stacking model 210 is deployed, it adjusts itself based on real-time data fed back from the stacking mechanical unit 300. The stacking model 210 continuously learns and updates to adapt to new stacking conditions, thereby ensuring the safe and stable loading of goods throughout the transportation process.

[0047] The stacking mechanical unit 300 executes the stacking task based on the stacking scheme sent by the processing unit 200 and the movement path of the non-standard goods 110. During the stacking process, the stacking model 210 adjusts the stacking scheme based on the geometric characteristics, vibration characteristics, weight characteristics, and offset characteristics of the goods fed back by the stacking mechanical unit 300. This invention can monitor the stacking of non-standard goods 110 and update the stacking scheme in real time to stack special goods reasonably and improve stacking efficiency. Specifically, this invention can automatically adjust the stacking layout, optimize the position of goods, reduce movement and collisions, and ensure the safe and stable loading of goods in the container, thereby improving the efficiency and reliability of the entire transportation process.

[0048] During the stacking process, the stacking model 210 adjusts the stacking scheme P based on the geometric features G, vibration features V, weight features W, and offset features O of the cargo fed back by the stacking mechanical unit 300. This invention can monitor the stacking of non-standard cargo 110 and update the stacking scheme P in real time to ensure reasonable stacking of special cargo and improve stacking efficiency. Specifically, this invention can automatically adjust the stacking layout, optimize cargo positions, reduce movement and collisions, and ensure the safe and stable loading of cargo within the container, thereby improving the efficiency and reliability of the entire transportation process.

[0049] P2=f(W, O; P1)=P1+α·△W+β·△O

[0050] P2 represents the new stacking scheme, and P1 represents the original stacking scheme. △W represents the adjustment amount calculated from the difference between the desired weight distribution and the actual weight distribution. △O represents the adjustment amount required based on the offset risk assessment. α and β are adjustment coefficients, corresponding to the adjustment coefficients for weight distribution and offset characteristic sensitivity, respectively. Preferably, stacking model 210 can evaluate △W by calculating the deviation of the center of gravity position. Stacking model 210 can evaluate △O by analyzing the displacement or tilting that may occur during handling or transportation.

[0051] In the same way, P2=f(G, V, W, O; P1)= P1+α·△W+β·△O+γ·△G+δ·△V.

[0052] △G represents the adjustment requirements based on the geometry of the cargo, such as to meet specific space utilization requirements or ensure proper spacing between cargo. △V represents the adjustments that must be made based on the vibration characteristics the cargo may encounter, aiming to reduce potential damage caused by vibration during transportation. γ and δ are adjustment coefficients for sensitivity to geometric and vibration characteristics, respectively.

[0053] like Figure 1 As shown, before stacking, the stacking model 210 adjusts the stacking scheme based on the spatial image data of the container collected by the image acquisition unit 100, and determines the stacking influence relationship between the stacked goods and non-standard goods 110 based on the type of the stacked goods and their transportation marking information, thereby adjusting the stacking parameters of the non-standard goods 110. This invention can intelligently adjust the stacking parameters of non-standard goods 110 according to the characteristics and interrelationships of different goods to achieve the optimal stacking layout.

[0054] Before the stacking process, the stacking model 210 first acquires spatial image data of the container from the image acquisition unit 100. Based on the spatial image data, the stacking model 210 performs a detailed spatial analysis of the container's interior.

[0055] For example, suppose a shipping container is to be loaded with electrical appliances, furniture, and several boxes of irregularly shaped artwork. Image acquisition unit 100 first scans the interior of the container, collecting precise image data about the internal spatial distribution. The container spatial image data acquired by image acquisition unit 100 is sent to stacking model 210. Stacking model 210 analyzes the spatial image data and the image data of the artwork, initially generating a standard stacking scheme for the appliances and furniture. At this stage, stacking model 210 pays particular attention to the types of goods already in the container and their marking information (such as size, weight, and stacking restriction labels), ensuring that these goods are arranged in positions best suited to their characteristics.

[0056] Subsequently, the stacking model 210 analyzes the shape, weight, and potential secure packaging requirements of the artworks to determine their spatial and physical impact on the already stacked goods. For example, if the artwork is very fragile, the stacking scheme generated by the stacking model 210 will determine that the artwork will not be placed under heavy objects that may move or tilt.

[0057] In order to adjust the stacking parameters of non-standard cargo 110, the stacking model 210 initially determined to place a larger but lighter artwork on the upper layer and a smaller but heavier sculpture in the center of the lower layer, in order to take advantage of the structural strength of the container and reduce the impact of vibration during transportation.

[0058] During the stacking process, there is a specific area or point known as the stacking core location. The stacking core location refers to the optimal location selected during the stacking process for placing goods to ensure the stability and security of the entire stacking structure.

[0059] The methods for determining the core location of the stacking structure include: receiving shipping marking information for the goods and understanding their physical characteristics, including size, weight, shape, and potential fragility. Based on the size and shape of each item, determine the geometry of the stack, such as rectangular, square, or other shapes. Calculate the center of gravity of the entire stacking structure; this helps determine the core location to maintain balance and stability. Select a stable foundation at the bottom of the stack; this could be a solid ground or a pallet. Select one or more items as the core; these items typically have significant weight and / or size and can provide support for other items. Vibrations and impacts may occur during transport; the core location should be chosen to minimize the impact of these factors on stack stability. Stacking model 210 selects the core location based on the principle that the center of gravity of the goods is located at the center of the stacking structure, or in a location that provides optimal support and balance.

[0060] For example, the coordinates of the center of gravity of the cargo are H(x) H , y H , z HThe center coordinates of the stacked structure are C(x). C , y C , z C ); the coordinates of the core placement position are Q(x) Q , y Q , z Q ).

[0061] The distance between the center of gravity and the center of the structure is: .

[0062] If the center of gravity (transportation marking information H) can be directly located above the structural center (transportation marking information C), then the transportation marking information Q = the transportation marking information C.

[0063] If the center of gravity (transportation mark information H) cannot be directly located above the structural center (transportation mark information C), then an optimal position (transportation mark information Q) needs to be calculated, such that the sum of the distance from transportation mark information H to transportation mark information Q and the distance from transportation mark information Q to transportation mark information C is minimized.

[0064] The core location for optimized stacking is: .

[0065] argmin means finding the parameter value that minimizes the function.

[0066] If the transport tag information d(H, C) is less than a certain preset threshold, then (the transport tag information Q = C).

[0067] If the transport label information d(H, C) is greater than the threshold, then an optimization algorithm (such as gradient descent, genetic algorithm, etc.) is used to find the optimal transport label information (Q).

[0068] like Figure 1 and Figure 5 As shown, during the stacking process, the stacking model 210 verifies the stability of the non-standard goods 110 at the stacking core position based on the geometric characteristics, vibration characteristics, weight characteristics, and offset characteristics of the non-standard goods 110 adjacent to the stacking core position.

[0069] The stacking model 210 adjusts the stacking scheme in real time based on the stability of the non-standard goods 110 at the core stacking location. By analyzing these characteristics and calculating the stability of the stacking pile 120, this invention can provide timely feedback and guidance during the actual stacking process to ensure the proper placement of goods and the stability of the stacking pile 120, thereby improving loading efficiency and safety.

[0070] Assuming there is a batch of orange crates (non-standard goods 110), the stacking model 210 collects the dimensional data of each crate using its own installed scanning equipment. The stacking model 210 uses a vibration sensor 330 to collect vibration data. A weight sensor 320 sends the weight data of the orange crates and other non-standard goods 110 to the stacking model 210. A position sensor monitors the offset of the goods during handling and sends this data to the stacking model 210.

[0071] The stacking model 210 analyzes the geometric and weight characteristics of the citrus crates to predict their stability during stacking. If it finds that some crates may cause stacking instability due to irregular shapes or uneven weight distribution, the stacking model 210 will adjust their placement scheme.

[0072] Stacking model 210 combines vibration and weight characteristics to evaluate the stability of easily vibrating items. Based on vibration parameters during handling and corresponding thresholds, stacking model 210 improves stability by adjusting the item's position or adding cushioning material. The thresholds corresponding to the vibration parameters are preset based on the item type. These thresholds can be obtained from a database based on the item type.

[0073] Based on the above analysis, the stacking model 210 generates adjustment plans in real time. For example, for citrus crates, the stacking model 210 may recommend using staggered stacking to increase stability; for items prone to vibration, it may suggest placing them in the most stable area of ​​the container and surrounding them with cushioning material.

[0074] like Figure 2 and Figure 5 As shown, after stacking, the stacking model 210 calculates the stability parameters of the goods based on the image data of the stacking pile 120 collected by the image acquisition unit 100, and performs similarity calculation between the image data of the stacking pile 120 and the image samples of stacking pile collapse cases to determine the safety of the stacking pile 120.

[0075] The stacking model 210 uses image data of the stacked pile collected by the image acquisition unit 100 to comprehensively assess the stability of the goods. By calculating the stability parameters of the goods and comparing them with image samples of existing stacked pile collapse cases, potential safety hazards can be detected in a timely manner, thereby taking necessary measures to ensure the safety and stability of the stacked pile 120.

[0076] like Figure 2 and Figure 5 As shown, the stacking model 210 determines the core position of each stacking state in the stacking scheme. During the stacking process, the stacking model 210 extracts the placement features of the goods at the core position of the stacking acquired by the image acquisition unit 100, and calculates the stability of the stacking pile 120 based on the placement features of the goods at the core position of the stacking.

[0077] The stability is calculated using the stability function as: S = α·center_of_mass_Score(F) + β·distribution_uniformity_Score(F) + γ·structural_integrity_Score(F).

[0078] `center_of_mass_Score(F)` calculates the center of gravity of the goods, `distribution_uniformity_Score(F)` evaluates the uniformity of the goods distribution, and `structural_integrity_Score(F)` evaluates the integrity of the stacking structure. α, β, and γ are weighting coefficients used to adjust the degree of influence of each factor on stability. F represents the placement feature vector (x...). F , y F , z F ).

[0079] .

[0080] d(H, C) is the Euclidean distance between the center of gravity of the cargo (H) and the center of the stacking structure (C).

[0081] .

[0082] .

[0083] By dynamically adjusting the stacking scheme based on real-time data, this invention can respond promptly to changes in cargo placement and the impact of the external environment, thereby ensuring the smooth progress of the entire stacking process and the final loading safety.

[0084] like Figure 3 and Figure 6 As shown, the processing unit 200 also includes a behavior recognition model 220. After the stacking is completed, the behavior recognition model 220 identifies the intervention behavior of the person 500 on the stacking pile 120 during the manual inspection process based on the image data collected by the image acquisition unit 100. The behavior recognition model 220 sends intervention behavior tags to the stacking model 210, so that the stacking model 210 associates the image data of the stacking scheme and intervention time with the intervention behavior tags. During the manual inspection process, by monitoring and recognizing various intervention behaviors of the person 500 in real time, such as rearranging goods and moving the position of goods, this invention can provide important information for subsequent stacking optimization and can also provide an optimization basis for optimizing the stacking model 210.

[0085] like Figure 3 and Figure 6As shown, the placement model 210 extracts corresponding image data based on the intervention behavior markers sent by the behavior recognition model 220, and compares the placement parameters of the intervention positions before and after the intervention based on the extracted image data corresponding to the intervention behavior markers. Figure 3 As shown, the placement model 210 sends placement parameters and image data of the intervention location before and after the intervention to the terminal 400 for manual confirmation. If the terminal 400 provides feedback indicating the intervention is effective, the placement model 210 optimizes the placement scheme based on the placement parameters of the intervention location after the intervention. The terminal 400 is a portable mobile terminal used by staff. The portable mobile terminal is equipped with a module or port capable of communicating with the placement model 210, allowing staff to confirm information or input manual suggestions through the terminal 400's screen, buttons, and other interactive components.

[0086] By sending the placement parameters and image data before and after the intervention to the terminal 400 for manual confirmation, and based on the effective intervention information fed back by the terminal 400, the placement model 210 can effectively process the intervention and optimize the placement scheme, ensuring that the intervention behavior can be identified and processed in a timely manner, and improving the automation and efficiency of the placement process.

[0087] According to a preferred embodiment, the stacking model 210 updates the stacking core position in real time based on the state of the stacking pile 120 of non-standard goods 110.

[0088] By continuously monitoring and analyzing the stacking status and stability of non-standard goods 110, the stacking model 210 can dynamically adjust the core position of the stacking to adapt to changes in the placement of the goods. This ensures that the system can update and optimize the core position in a timely manner during the stacking process, thereby improving the rationality and stability of the stacking layout. Example 2

[0089] This invention provides a method for monitoring the stacking of containers, such as... Figure 7 As shown, the method includes:

[0090] S1: Collect stacking image data corresponding to non-standard goods 110.

[0091] S2: Construct a stacking model 210 corresponding to non-standard goods 110 based on stacking image data and artificial intelligence model.

[0092] S3: Adjust the movement path of non-standard goods 110 based on the stacking scheme generated by the stacking model 210. Preferably, during the stacking process, the stacking scheme is adjusted based on the geometric characteristics, vibration characteristics, weight characteristics, and offset characteristics of the goods.

[0093] S4: Execute the stacking task based on the stacking scheme sent by the processing unit 200 and the movement path of the non-standard goods 110;

[0094] By collecting stacking image data corresponding to non-standard cargo 110, a comprehensive perception of the container's internal layout is achieved. This allows for accurate acquisition of the position, status, and layout of each cargo within the container, providing necessary data support for the subsequent stacking process. Utilizing stacking image data and an artificial intelligence model, a stacking model 210 corresponding to non-standard cargo 110 is constructed. Based on image data and deep learning model training, the model accurately identifies the type, shape, and position of non-standard cargo 110 and establishes the corresponding stacking model 210, providing a reliable basis for the generation and optimization of subsequent stacking schemes. By adjusting the movement path of non-standard cargo 110, dynamic optimization of the stacking scheme is achieved. The position and layout of cargo can be flexibly adjusted according to the geometric, vibration, weight, and offset characteristics of different cargoes to maximize the loading efficiency and stability of the container. This invention comprehensively considers multiple cargo characteristics and dynamically optimizes the stacking scheme to ensure stable cargo placement and safe container loading.

[0095] S5: Adjust the stacking scheme based on the spatial image data of the container, and determine the stacking influence relationship between the stacked goods and non-standard goods 110 based on the type of stacked goods and their transport marking information, thereby adjusting the stacking parameters of non-standard goods 110. This invention intelligently adjusts the stacking scheme according to the actual internal space of the container, improving the efficiency and accuracy of the stacking process.

[0096] It should be noted that the specific embodiments described above are exemplary. Those skilled in the art can devise various solutions inspired by the disclosure of this invention, and these solutions all fall within the scope of this invention and its protection. Those skilled in the art should understand that this specification and its accompanying drawings are illustrative and not intended to limit the scope of the claims. The scope of protection of this invention is defined by the claims and their equivalents. This specification contains multiple inventive concepts; phrases such as "preferredly" or "according to a preferred embodiment" indicate that the corresponding paragraph discloses an independent concept. The applicant reserves the right to file divisional applications based on each inventive concept.

Claims

1. A container stacking monitoring system, characterized in that, include: Image acquisition unit (100): Acquires stacking image data corresponding to non-standard goods (110); Processing unit (200): Constructs a stacking model (210) corresponding to the non-standard goods (110) based on the stacking image data and artificial intelligence model; adjusts the movement path of the non-standard goods (110) based on the stacking scheme generated by the stacking model (210); Stacking machine unit (300): performs stacking tasks based on the stacking scheme sent by the processing unit (200) and the movement path of the non-standard goods (110); During the stacking process, the stacking model (210) adjusts the stacking scheme based on the geometric characteristics, vibration characteristics, weight characteristics and offset characteristics of the goods fed back by the stacking mechanical unit (300).

2. The container stacking monitoring system according to claim 1, characterized in that, Before stacking, the stacking model (210) adjusts the stacking scheme according to the spatial image data of the container collected by the image acquisition unit (100), and determines the stacking influence relationship between the stacked goods and the non-standard goods (110) according to the type of the stacked goods and their transportation marking information, thereby adjusting the stacking parameters of the non-standard goods (110).

3. The container stacking monitoring system according to claim 2, characterized in that, After stacking, the stacking model (210) calculates the stability parameters of the goods based on the image data of the stack (120) collected by the image acquisition unit (100), and performs similarity calculation between the image data of the stack (120) and the image samples of stack collapse cases to determine the safety of the stack (120).

4. The container stacking monitoring system according to claim 3, characterized in that, The stacking model (210) determines the core position of each stacking state in the stacking scheme. During the stacking process, the stacking model (210) extracts the placement features of the goods at the core stacking location collected by the image acquisition unit (100), and calculates the stability of the stacking pile (120) based on the placement features of the goods at the core stacking location.

5. The container stacking monitoring system according to any one of claims 1 to 4, characterized in that, During the stacking process, the stacking model (210) verifies the stability of the non-standard goods (110) at the stacking core position based on the geometric characteristics, vibration characteristics, weight characteristics and offset characteristics of the non-standard goods (110) adjacent to the stacking core position, and adjusts the stacking scheme in real time according to the stability of the non-standard goods (110) at the stacking core position.

6. The container stacking monitoring system according to any one of claims 1 to 4, characterized in that, The processing unit (200) also includes a behavior recognition model (220). After the stacking is completed, the behavior recognition model (220) identifies the intervention behavior of the stacking pile (120) during the manual inspection process based on the image data collected by the image acquisition unit (100). The behavior recognition model (220) sends an intervention behavior tag to the scrambling model (210) so that the scrambling model (210) associates the image data of the scrambling scheme and intervention time with the intervention behavior tag.

7. The container stacking monitoring system according to any one of claims 1 to 4, characterized in that, The grading model (210) extracts corresponding image data based on the intervention behavior markers sent by the behavior recognition model (220), and compares the grading parameters of the intervention positions before and after the intervention based on the extracted image data corresponding to the intervention behavior markers. The placement model (210) sends the placement parameters and image data of the intervention position before and after the intervention to the terminal (400) for manual confirmation. When the terminal (400) provides feedback that the intervention is effective, the scrambling model (210) optimizes the scrambling scheme based on the scrambling parameters of the intervention position after the intervention.

8. The container stacking monitoring system according to any one of claims 1 to 4, characterized in that, The stacking model (210) updates the core position of the stacking in real time based on the status of the stacking pile (120) of non-standard goods (110).

9. A method for monitoring the stacking of containers, characterized in that, The method includes: Collect stacking image data corresponding to non-standard goods (110); Based on the stacking image data and artificial intelligence model, a stacking model (210) corresponding to the non-standard goods (110) is constructed. The movement path of the non-standard goods (110) is adjusted based on the stacking scheme generated by the stacking model (210); The stacking task is performed based on the stacking scheme sent by the processing unit (200) and the movement path of the non-standard goods (110); During the stacking process, the stacking scheme is adjusted based on the geometric characteristics, vibration characteristics, weight characteristics, and offset characteristics of the goods.

10. The container stacking monitoring method according to claim 9, characterized in that, The method further includes: The stacking scheme is adjusted based on the spatial image data of the container, and the stacking influence relationship between the stacked goods and the non-standard goods (110) is determined based on the type of the stacked goods and their transport marking information, thereby adjusting the stacking parameters of the non-standard goods (110).

Images