Visual intelligent system for material transfer and its construction method

Abstract

The present invention discloses a visualization intelligent system for material transportation and a construction method thereof. The construction method includes: building a three-dimensional framework; building a model; building 3D modeling and a visualization scene, thus forming a visualization intelligent system for material transportation, which includes a task module, a manual input module, a material scanning module, a plan download module, an intelligent generation module, and a plan viewing module. According to information such as the specifications of transportation tools and the shape, size, quantity, weight, etc. of goods, the present invention can determine a feasible loading method, so that under the conditions of meeting constraints such as given volume constraints, load-bearing capacity constraints, stability constraints, loading sequence, etc., the total volume of the items contained in the container is as large as possible, that is to say, the filling rate of the container is as large as possible, thereby more effectively utilizing the space of the transportation container during transportation and reducing transportation costs.

Description

Technical Field

[0001] The present invention belongs to the technical field of material transportation, and particularly relates to a visual intelligent system for material transportation and a construction method thereof. Background Technique

[0002] The current military forces are facing huge challenges in material transportation:

[0003] (1) The requirement of full-territory operations requires that support equipment can be quickly mobilized, transferred, and loaded, and higher requirements are put forward for the rapid inventory and inspection of materials.

[0004] (2) There are a wide variety of loaded materials and the order placement is complex. It is not easy to find specific materials during unloading, and there is no effective means to search for lost tools. Only a large amount of manpower and time can be spent to search slowly, and the efficiency is extremely low.

[0005] (3) Material tags cannot meet the current material positioning and control. Material tags are easily damaged, dropped, and polluted; they cannot transmit material information in a narrow space, making it impossible for relevant personnel to effectively locate and collect.

[0006] (4) The material management work cannot be assigned to specific individuals. When materials are lost, it is easy for operators to shift responsibilities to each other. The imperfect material management responsibility system directly leads to low utilization rate of material loading and unloading, resulting in problems such as asset losses, tool losses, and safety hazards.

[0007] The traditional method for the military to transport materials is the manual experience operation method (as the name implies), and the existing logistics platforms related to transportation in the market do not have the three functions of visualization, intelligent generation, and human-computer interaction and modification (or only meet one of them), which all affect the efficiency in the process of material transportation work. Summary of the Invention

[0008] In view of the above existing problems, the present invention provides a construction method for a visual intelligent system for material transportation, which solves the problem of large-scale material transportation and quickly, systematically, and automatically generates a transportation plan for large-scale materials.

[0009] The technical solution adopted by the present invention is:

[0010] A construction method for a visual intelligent system for material transportation, comprising:

[0011] Building a three-dimensional framework: The front-end main body uses the Vue.js framework, and on this basis, the WebGL function encapsulated by the Three.js framework is used to implement the three-dimensional part of the main body;

[0012] Model: The object model is built using the Blender modeling software, and the Three.js loader is used to load the model. The text / binary model file is converted into a Three.js object structure;

[0013] 3D Modeling and Visualization Scene Setup:

[0014] The home.vue file is created in the page folder under the wuzi project file. The page layout framework is built using various components, specifically: the view component, text component, block component, and input component;

[0015] First, an array containing multiple objects is created in the database. At the top of the page, the block component is used to display different elements in the object through the index of each array object, including the task module, manual input module, scanned materials module, solution download module, intelligent generation module, and solution viewing module;

[0016] The interface layouts of the task module, manual input module, scanned materials module, solution download module, intelligent generation module, and solution viewing module are built respectively. By writing the clickModule(index) function for the @click click event in the method in the view, the current index is changed to the index of the currently clicked item. Then, by judging whether the current index is the index of its own object in different views and obtaining the judgment result through v-show, the function of switching different interfaces is realized;

[0017] After completing the page setup, the functions are constructed in the method, and the functions are uniformly written in the.js file. By storing the.js file in each module, the functions can be called.

[0018] Preferably, in the construction of the interface layout, different rendering styles in the imported free.css file are used for page rendering through the class in the view component.

[0019] Preferably, in the construction of the interface layout, the Python language is used to implement 3D animation modeling, including using the Matplotlib drawing library, NumPy tool, and 3D drawing library to implement visual images and three-dimensional shapes, and using the mayavi library to implement 3D dynamic visualization operations.

[0020] Material Transfer Visualization Intelligent System, including:

[0021] Task module, used to view the task list, create daily tasks, and determine the transportation vehicle;

[0022] Manual input module for inputting material data;

[0023] Material scanning module for scanning material data. After importing the material data, a material transfer plan is automatically generated;

[0024] Intelligent generation module for receiving the material transfer plan and presenting the material transfer method in the form of three views;

[0025] Plan viewing module for viewing historical plans and displaying the positions of materials in the carrier's cabin;

[0026] Plan downloading module for downloading and saving the material transfer plan, generating a spreadsheet plan list, and calibrating the determined coordinate positions of each material.

[0027] Preferably, a space planning model algorithm is stored in the material scanning module. The space planning model algorithm is programmed using VC++ 6.0. The data processing process is linked to the database, and a standard data communication interface is used to connect the database with the front-end uni-app of the system development platform.

[0028] Preferably, the space planning model algorithm includes the following steps:

[0029] Step 1: According to the spatial dimensions of the imported transportation tool, in meters; record the length as x, the width as y, and the height as h;

[0030] Step 1: Input the dimensions of each material. The lengths are respectively recorded as: x1 x2 x3 …… x n ; the widths are respectively recorded as: y1 y2 y3…… y n ; the heights are respectively recorded as: h, h2 h3 …… h n ; respectively take the maximum values of the three groups of numbers:

[0031] X = max(x1 x2 x3 …… x n )

[0032] Y = max(y1 y2 y3 …… y n )

[0033] H = max(h1 h2 h3 …… h n )

[0034] Divide the space of the transportation platform into (x / X)×(y / Y)×(h / H) sub-spaces, number them, starting from the origin, the grid space coordinates range from (0, 0, 0) to ((x / X), (y / Y), (h / H)), and the sequence number of each grid is (n) = h′·3 + y′·1 + x′·1, where x′y′h′ are the coordinates of each grid. The sequence numbers of each grid are arranged from large to small. It is stipulated that the greater the sequence number of the grid space, the greater the importance of the item placed in it;

[0035] Step 2: Calculate the importance of each item, that is, W. The importance of the item includes item attributes, center of gravity position, and item weight. Among them, the item attributes are given different importance scores through the expert scoring method; the function for calculating the importance of the item is obtained as:

[0036] W = α·W - β·m + Z

[0037] In the formula, the larger W is, the larger the sequence number of the corresponding spatial position is. w represents the item attribute score; m represents the item weight; α and β are constant values; Z is a correction and regulation coefficient, and its assignment conditions are: (1) For any item, the initial value is 0; (2) If the material plan generated for the first time does not meet the following constraint conditions, then introduce the Z value. The constraint conditions are:

[0038]

[0039] In the formula, x″y″h″ is the center of gravity position of the total materials, and X YH is the center position of the cabin;

[0040] Step 3: Generate the placement position of the materials in the platform space according to the calculated importance. The larger the W value is, the larger the sequence number of the corresponding spatial position is, that is, generate the loading plan;

[0041] Step 4: Perform height correction on the loading plan;

[0042] Step 5: Perform depth correction on the loading plan;

[0043] Step 6: Judge whether the corrected loading plan meets the constraint conditions. If it is satisfied, generate the final plan; otherwise, introduce the adjustment and correction coefficient and return to Step 2.

[0044] Preferably, in Step 2, the values of α and β are obtained as follows:

[0045] Adopt the BP neural network model. First, set the objective function: In the formula, ∑v iLet \(V_{total}\) represent the total volume of the loaded materials, \(V\) represent the total transportation space, and \(\mu\) represent the space utilization rate. The input layer of the BP neural network model is represented by the vector \((\alpha, \beta, Z, w, m)\), the output layer is \((\mu)\), and the weight set of the hidden layer is self-repaired and updated according to the Hebb rule. Three hidden layers are set for perception training, and the output result is obtained. End condition: \(\mu\geq90\%\); record the values of \(\alpha\) and \(\beta\).

[0046] Preferably, in step 4, the height correction algorithm for the loading plan is as follows:

[0047] (1) Determine whether all sub-spaces are marked. If there are unmarked sub-spaces, continue to step (2);

[0048] (2) Select an unmarked sub-space and determine whether the \(h'\) coordinate of this sub-space is greater than 1. If it is satisfied, continue to step (3); otherwise, go to step (1).

[0049] (3) Search for the sub-space in the marked sub-spaces where the \(x'\) and \(y'\) coordinates are equal to those of this sub-space and the \(H\) coordinate is \(h' - 1\). If there is a sub-space that meets this condition, move the material corresponding to this sub-space down one position and place it on the sub-space with coordinates \((x', y', h' - 1)\), and mark this sub-space. Then return to step (1) until all sub-spaces are marked and end.

[0050] Preferably, in step 4, the depth correction algorithm for the loading plan is as follows:

[0051] (1) Determine whether all sub-spaces are marked. If there are unmarked sub-spaces, continue to step (2);

[0052] (2) Select an unmarked sub-space and determine whether the \(x'\) coordinate of this sub-space is greater than 1. If it is satisfied, continue to step (3); otherwise, go to step (1).

[0053] (3) Search for the sub-space in the marked sub-spaces where the \(y'\) and \(h'\) coordinates are equal to those of this sub-space and the \(X\) coordinate is \(x' - 1\). If there is a sub-space that meets this condition, move the material corresponding to this sub-space down one position and place it on the sub-space with coordinates \((x' - 1, y', h')\), and mark this sub-space. Then return to step (1) until all sub-spaces are marked and end.

[0054] Advantages of the present invention: According to the specifications of the means of transportation and information such as the shape, size, quantity, and weight of the goods, the present invention can determine a feasible loading method, so that under the conditions of meeting the given volume constraints, load-bearing capacity constraints, stability constraints, loading sequence, etc., the total volume of the items contained in the container is as large as possible, that is, the filling rate of the container is as large as possible, thereby more effectively utilizing the space of the transportation container during transportation and reducing transportation costs. Brief Description of the Drawings

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0056] Figure 1 Shown is the display interface of the manual input module in this embodiment;

[0057] Figure 2 Shown is the display interface of the intelligent generation module in this embodiment;

[0058] Figure 3 Shown is the display interface of the solution viewing module in this embodiment;

[0059] Figure 4 Shown is the flowchart of the spatial planning model algorithm in this embodiment;

[0060] Figure 5 Shown is the flowchart of the height correction of the material plan in this embodiment;

[0061] Figure 6 Shown is the flowchart of the depth correction of the material plan in this embodiment. Detailed Embodiments

[0062] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0063] Accordingly, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0064] This embodiment specifically provides a method for building a visualization intelligent system for material transfer, including:

[0065] Three-dimensional framework

[0066] The application front-end main body uses the Vue.js framework, and on this basis, uses the WebGL function encapsulated by the Three.js framework to implement the three-dimensional part of the main body, which can facilitate cross-platform use of the program. WebGL requires knowledge of computer graphics. If directly used, one must comprehensively understand shader syntax and write vertex shaders by oneself. Three.js solves the problem of complex WebGL development. It encapsulates concepts such as scene, camera, geometry, 3D model loader, light, material, shader, animation, particle, and mathematical tools. Using Three.js can conveniently skip these underlying layers, allowing developers to write graphics applications by writing traditional JavaScript.

[0067] Model

[0068] Build the object model through modeling software such as blender, and use the Three.js loader to achieve model loading, converting the text / binary model file into a Three.js object structure. Each loader understands a specific file format, such as JSON / OBJ / MTL / STL, etc. Json can be exported through blender or 3DsMax. They have their own exportjson plugins. After processing the model texture and animation in the software, export the json file and the corresponding texture file to the front end.

[0069] 3D modeling and construction of the visualization scene

[0070] The interface display and function implementation of the material visualization intelligent system are developed based on the front-end framework of uni-app. Uni-app is a front-end framework for developing cross-platform applications using Vue.js. By writing a set of code, it can be compiled to multiple platforms such as Android, iOS, H5, and mini programs. The material visualization intelligent system is finally compiled to the H5 platform and displayed on the web page. During the program writing process, the HTML5+CSS+JavaScript mode is adopted for development.

[0071] Based on the designed UI design draft, a home.vue file was created in the page folder under the wuzi project file to implement the page layout framework. The page was built using various components, including the view component to wrap the element content, the text component to write the text content, the block component to perform list rendering, and the input component to form an input box. First, we created an array containing multiple objects in data. At the top of the page, the block component was used to display different elements in the object through the index of each array object, including: task module, manual input module, material scanning module, solution download module, intelligent generation module, and solution viewing module, which formed a task bar. Different interface layouts were built within each module. By writing the clickModule(index) function for the @click click event in the view and in the method, the current index was changed to the index of the currently clicked item. Then, by judging whether the current index in different views was the index of its own object and obtaining the judgment result through v-show, the function of switching different interfaces was realized.

[0072] In the construction of the lower half of the page, different rendering styles from the imported free.css file were used in the class of the view component for page rendering. free.css is a custom rendering file that contains a large number of rendering styles, including layout styles, colors, fonts, spacing, click animations, etc. We can also obtain various icons we need and are satisfied with from the iconfont.css file, making the interface layout more beautiful and the effect more satisfactory. By continuously improving the.css file and adding more required rendering styles, the effects we need to achieve can be completed. At the same time, as the.css file is continuously improved, the construction of the next page will become simpler and more convenient. There are a large number of css libraries available for reference on the official website of uni-app and various online materials, and the documentation is very rich.

[0073] After completing the overall page setup, we still need to further improve the functions and algorithms implemented in the software. There are many buttons on the interface, including creating new data, scanning and importing data, generating spreadsheets, and loading scenarios. All of these involve data storage and export, page lifecycle, and the use of hooks such as page monitoring. It also involves writing functions to implement the functions of multiple button events. We can construct functions in the method. We can also write functions uniformly in the js file. We can then introduce the.js file in multiple pages to call the functions in it, making it more convenient to call some commonly used functions. At the same time, when a large number of data formats need to be called, to make the page display more smooth, we can store the data in the js file and call it through references.

[0074] In the wuzi uni-app project, we made good use of the framework advantages of uni-app, separated the vue, css, and js files, and completed the page setup and function implementation of the entire software system by calling and introducing them separately.

[0075] In terms of the implementation of 3D animation technology, we mainly use the Python language for programming. It has the characteristics of being easy to expand, a glue language, and open source. Here we mainly use the two three-dimensional visualization and 3D painting libraries, Matplotlib and NumPy, to implement visual images and three-dimensional shapes. And we use the mayavi library to implement 3D dynamic visualization operations.

[0076] 1. 3D Movement

[0077] translate3d(x, y, z) makes the element move in these three dimensions and can also be written separately.

[0078] translateX(x), translateY(y), translateZ(z).

[0079] transform: translateX(100px); / / Move along the X-axis

[0080] 2. 3D Scaling

[0081] scale3d(number, number, number) can also be written separately as scaleX(), scaleY(), scaleZ().

[0082] 3. 3D Rotation

[0083] rotate3D(x, y, z, angle) specifies the coordinate axes for rotation

[0084] rotateX(angle) rotates the element along the X-axis.

[0085] rotateY(angle) rotates the element along the Y-axis.

[0086] rotateZ(angle) rotates the element along the Z-axis.

[0087] This embodiment also provides a visualization intelligent system for material transfer, including:

[0088] A task module for viewing the task list, creating daily tasks, and determining the transportation vehicle;

[0089] A manual input module for inputting material data, such as Figure 1 shown;

[0090] A material scanning module for scanning material data. After importing the material data, a material transfer plan is automatically generated;

[0091] An intelligent generation module for receiving the material transfer plan and displaying the material transfer method in the form of three views, such as Figure 2 shown;

[0092] A plan viewing module for viewing historical plans and displaying the positions of materials in the transportation vehicle cabin, such as Figure 3 shown;

[0093] A plan downloading module for downloading and saving the material transfer plan, generating an electronic spreadsheet plan list, and calibrating the determined coordinate positions of each material.

[0094] A space planning model algorithm is stored in the material scanning module. This space planning model algorithm is programmed using VC++6.0. The data processing process is linked to the database, and a standard data communication interface is used to connect the database with the front-end uni-app of the system development platform.

[0095] The space planning model algorithm, as Figure 4 shown, includes the following steps:

[0096] Step 1: According to the spatial dimensions of the imported transportation vehicle, in meters; record the length as x, the width as y, and the height as h;

[0097] Step 1: Input the dimensions of each material. The lengths are respectively recorded as: x1 x2 x3 …… x n ; the widths are respectively recorded as: y1 y2 y3…… y n ; the heights are respectively recorded as: h1 h2 h3 …… h n ; respectively take the maximum values of the three groups of numbers:

[0098] X = max(x1, x2, x3, ……, x n )

[0099] Y = max(y1, y2, y3, ……, y n )

[0100] H = max(h1, h2, h3, ……, h n )

[0101] Divide the space of the transportation platform into (x / X)×(y / Y)×(h / H) sub - spaces and number them. Starting from the origin, the grid - space coordinates range from (0, 0, 0) to ((x / X), (y / Y), (h / H)). The sequence number of each grid is (n)=h′·3 + y′·1 + x′·1, where x′, y′, h′ are the coordinates of each grid. The sequence numbers of each grid are arranged from large to small. It is stipulated that the greater the sequence number of the grid space, the greater the importance of the item placed in it;

[0102] Step 2: Calculate the importance degree of each piece of material, that is, W. The importance degree of the material includes material attributes, center - of - gravity position, and material weight. Among them, the material attributes are given importance - degree scores for different materials through the expert scoring method. See the following table for details:

[0103] Material Attributes Score Weapons and Ammunition 5 Precision Machinery 4 Aviation Materials and Equipment 3 Pod Mounts 2 Carried Supplies 1 Others Customize

[0104] Center - of - gravity position (only used when the transportation tool is an aircraft)

[0105] Load trimming is carried out to ensure that the center of gravity of the total materials is near the center of the aircraft cabin. The calculation method of the total - materials center of gravity can be obtained through the center - of - gravity calculation formula with the center of gravity of each material (here it is considered that the center - of - gravity position of each material is at the center of the material geometry). The difference between the total - materials center - of - gravity position and the cabin center cannot exceed 1 meter.

[0106] Weight. The greater the weight, the more inward the placement position of the material, that is, the smaller the sequence number of the corresponding space position.

[0107] The function for calculating the importance degree of the material is obtained as:

[0108] W = α·w - β·m+z

[0109] In the formula, the greater W is, the greater the sequence number of the corresponding space position. w represents the material - attribute score; m represents the material weight; α and β are constant values; z is a correction and regulation coefficient, and its assignment conditions are: (1) For any piece of material, the initial value is 0; (2) If the material plan generated for the first time does not meet the following constraint conditions, then introduce the z value, and it is stipulated as:

[0110] For the most important material, the corresponding z value is 1. For the second most important material, the z value is 2, and so on. That is, the z value is the sequence number of the material importance ranking.

[0111] The constraint conditions are as follows:

[0112]

[0113] In the formula, x″y″h″ is the centroid position of the total materials, and X YH is the center position of the compartment;

[0114] Step 3: Generate the placement position of the materials in the platform space according to the calculated importance. The larger the W value, the larger the sequence number of the corresponding space position, that is, generate the loading plan;

[0115] Step 4: Perform height correction on the loading plan;

[0116] Step 5: Perform depth correction on the loading plan;

[0117] Step 6: Determine whether the corrected loading plan meets the constraint conditions. If it is satisfied, generate the final plan; otherwise, introduce an adjustment correction coefficient and return to Step 2.

[0118] The value-taking process of α and β is as follows:

[0119] For the value-taking of α and β, the simple expert scoring method is no longer adopted here. Because the material types are different, it is difficult to use a constant value to uniformly measure. The weights involved in the inherent attributes and weights of the materials in different situations are different. For example, in the case of using trains as the main means of transportation, the attributes of the materials can better determine the importance of the materials themselves. However, in the case of air transportation, because the load balance of the aircraft needs to be considered, the weight of the materials becomes more important.

[0120] To solve this problem, I break out of the traditional method field and adopt a neural network algorithm with strong adaptability. The neural network algorithm performs massive calculations using big data for different environments, and finds the optimal values suitable for different scenarios through iterations. Computer technology can make this algorithm a reality.

[0121] Adopt the BP neural network model. First, set the objective function: In the formula, ∑v iLet \(V_{total}\) represent the total volume of the loaded materials, \(V\) represent the total transportation space, and \(\mu\) represent the space utilization rate. The input layer of the BP neural network model is represented by the vector \((\alpha, \beta, z, w, m)\), the output layer is \((\mu)\), and the weight set of the hidden layer is self-repaired and updated according to the Hebb rule. Three hidden layers are set for perception training, and the output result is obtained. The end condition is: \(\mu\geq90\%\). Record the values of \(\alpha\) and \(\beta\). Based on these values, after calculating the space where each material is located, it may occur that the position of some materials is in a suspended state, which does not conform to the actual situation. Therefore, the above calculation process needs to be corrected once more.

[0122] The height correction algorithm for the loading plan is as Figure 5 shown, including:

[0123] (1) Judge whether all sub-spaces are marked. If there are unmarked sub-spaces, continue to step (2);

[0124] (2) Select an unmarked sub-space and judge whether the \(h'\) coordinate of this sub-space is \(\gt1\). If it is satisfied, continue to step (3); otherwise, go to step (1).

[0125] (3) Search for the sub-space in the marked sub-spaces whose \(x'\) and \(y'\) coordinates are equal to those of this sub-space at the same time, and the \(H\) coordinate is \(h' - 1\). If there is a sub-space that meets this condition, then adjust the material corresponding to this sub-space down by one position and place it on the sub-space with the coordinates \((x', y', h' - 1)\), and mark this sub-space. Then return to step (1) until all sub-spaces are marked and end.

[0126] The depth correction algorithm for the loading plan is as Figure 6 shown, including:

[0127] (1) Judge whether all sub-spaces are marked. If there are unmarked sub-spaces, continue to step (2);

[0128] (2) Select an unmarked sub-space and judge whether the \(x'\) coordinate of this sub-space is \(\gt1\). If it is satisfied, continue to step (3); otherwise, go to step (1).

[0129] (3) Search for the sub-space in the marked sub-spaces whose \(y'\) and \(h'\) coordinates are equal to those of this sub-space at the same time, and the \(X\) coordinate is \(x' - 1\). If there is a sub-space that meets this condition, then adjust the material corresponding to this sub-space down by one position and place it on the sub-space with the coordinates \((x' - 1, y', h')\), and mark this sub-space. Then return to step (1) until all sub-spaces are marked and end.

[0130] According to the specifications of the transportation vehicle and information such as the shape, size, quantity, weight, etc. of the goods, the present invention can determine a feasible loading method, so that under the conditions of meeting given constraints such as volume constraint, load-bearing capacity constraint, stability constraint, loading sequence, etc., the total volume of the items contained in the container is as large as possible, that is to say, the filling rate of the containerized container is as large as possible, thereby more effectively utilizing the space of the transportation container during transportation and reducing transportation costs.

[0131] As described above, it is only used to illustrate the technical solution of the present invention rather than to limit it. Other modifications or equivalent replacements made by those of ordinary skill in the art to the technical solution of the present invention shall be covered within the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. Visual intelligent system for material transfer, characterized in that Including: A task module for viewing a task list, creating a daily task, and determining a vehicle; A manual input module for inputting material data; A material scanning module for scanning material data and automatically generating a material transfer plan after importing the material data; An intelligent generation module for receiving the material transfer plan and displaying the material transfer method in the form of three views; A plan viewing module for viewing historical plans and showing the positions of materials in the vehicle cabin; A plan downloading module for downloading and saving the material transfer plan, generating an electronic spreadsheet plan list, and calibrating the determined coordinate positions of each material; The material scanning module stores a space planning model algorithm, which is programmed using VC++ 6.0, the data processing process is linked to the database, and a standard data communication interface is used to connect the database and the front-end uni-app of the system development platform; The space planning model algorithm includes the following steps: Step 1: According to the spatial dimensions of the imported means of transportation, in meters; record the length as x , the width as y , and the height as h ; Step 2: Input the dimensions of each material. The lengths are denoted as: ; The widths are denoted as: ; The heights are respectively denoted as: ; Take the maximum values of the three groups of numbers respectively: Divide the space of the transportation platform into sub-spaces and label them. Starting from the origin, the grid space coordinates are to , and the sequence number of each grid is which is the coordinate of each grid. The sequence numbers of each grid are arranged from largest to smallest. It is stipulated that the greater the sequence number of the grid space, the greater the importance of the items placed in it; Step 3: Calculate the importance degree of each material, that is , the importance degree of the material includes material attributes, center of gravity position, and material weight. Among them, the importance degree score of different materials is assigned to the material attributes by the expert scoring method; Wherein, The larger it is, the larger the sequence number of the corresponding spatial position is. represents the material attribute score; represents the material weight; is a constant value; is a correction control coefficient, and its assignment conditions are: (1) for any piece of material, the initial value is 0; (2) if the material plan generated for the first time does not meet the following constraint conditions, then introduce value, and the constraint conditions are: In the formula, is the coordinate value of the centroid position of the total materials, is the coordinate value of the central position of the compartment; Step 4: Generate the placement positions of materials in the platform space according to the calculated importance levels. The larger the value, the larger the sequence number of the corresponding space position, that is, a loading plan is generated. Step 5: Perform height correction on the loading plan; Step 6: Perform depth correction on the loading plan; Step 7: Determine whether the corrected loading plan meets the constraint conditions. If it is satisfied, generate the final plan; otherwise, introduce a correction control coefficient and return to Step 3.

2. The visualized intelligent system for material transfer according to claim 1, wherein In step 3, is obtained as follows: Adopt a BP neural network model and first set the objective function: , where represents the total volume of the loaded materials, represents the total transportation space, represents the space utilization rate; the input layer of the BP neural network model is represented by a vector as , the output layer is , the hidden layer weight set is self-repaired and updated according to the Hebb rule. Three hidden layers are set for perception training, and the output result is obtained. End condition: ; Record value.

3. The visualized intelligent material transfer system according to claim 1, wherein, In Step 5, the loading plan height correction algorithm: (1) Determine whether all sub-spaces are marked. If there are unmarked sub-spaces, continue with Step (2); (2) Select the unlabeled subspace and determine the coordinate >

1. If satisfied, continue with step (3); otherwise, go to step (1) (3) Search for the subspace in the labeled subspaces whose coordinates are equal at the same time and the height coordinate is If there is a subspace that meets this condition, move the supplies corresponding to this subspace down one position to the coordinates on a subspace of, mark this subspace, and return to step (1) until all subspaces are marked, and then end.

4. The visualized intelligent system for material transfer according to claim 1, wherein In Step 6, the loading plan depth correction algorithm: (1) Determine whether all sub-spaces are marked. If there are unmarked sub-spaces, continue with Step (2); (2) Select the unlabeled subspace and determine the coordinates >

1. If satisfied, proceed to step (3); otherwise, go to step (1). (3) Search for the subspace in the labeled subspaces where the , , and the coordinates are equal at the same time, and the length coordinate is . If there is a subspace that meets this condition, move the material corresponding to this subspace down one position and place it on the subspace with the coordinate . Then mark this subspace and return to step (1) until all subspaces are marked and end.

Images