Mechanical exterior decoration model construction method and system based on neural network

Abstract

The invention discloses a mechanical exterior model construction method and system based on a neural network. The method comprises the following steps: reading a product standard image from a product image database according to PEST category data and product semantic data; extracting brand element features from a brand element database according to the brand semantic data; inputting the product standard image into a preset exterior part identification model to obtain a shape identification feature and a color identification feature of the exterior part; generating an exterior part model according to the shape recognition features of the exterior part, and generating a color matching scheme according to the color recognition features; the exterior trimming part model is called to build a mechanical exterior trimming part basic model, the brand element features are loaded into the mechanical exterior trimming part basic model, and a color matching scheme is called to render the mechanical exterior trimming part basic model to obtain a mechanical exterior trimming part application model; according to the method, designers are assisted in producing the engineering machinery exterior trimming parts meeting brand semantics, product semantics and market demand style types.

Description

Technical Field

[0001] The present invention belongs to the technical field of exterior decoration evaluation, and in particular relates to a method and system for constructing a mechanical exterior decoration model based on a neural network. Background Art

[0002] With the acceleration of industrialization and the vigorous development of the construction industry, the appearance design and exterior decoration configuration of construction machinery are receiving more and more attention. The exterior decoration of construction machinery is not only related to the visual image of the product, but also directly affects the market competitiveness and brand value of the product.

[0003] The existing construction process of engineering machinery exterior accessories mainly focuses on a single design link, and only simply considers the practicality of mechanical exterior accessories. However, the following problems still exist in actual operation: 1) Brand semantics and product semantics are not effectively processed and analyzed, resulting in insufficient brand richness of engineering machinery exterior accessories; 2) The acquired market data is not used for more detailed model construction and trend analysis, resulting in the production of engineering machinery exterior accessories that do not meet market demand. Summary of the invention

[0004] The present invention provides a method and system for constructing a mechanical exterior decoration model based on a neural network, which assists designers in producing engineering machinery exterior decoration parts that meet brand semantics, product semantics and market demand style types.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: A first aspect of the present invention provides a method for constructing a mechanical exterior model based on a neural network, comprising: Acquire brand semantic data and PEST category data and product semantic data of mechanical exterior parts; read product standard images from a product image database based on the PEST category data and product semantic data; extract brand element features from a brand element database based on the brand semantic data; Input the product standard image into the preset exterior component recognition model to obtain the shape recognition features and color recognition features of the exterior component; generate the exterior component model according to the shape recognition features of the exterior component, and generate the color matching scheme according to the color recognition features; store the exterior component model and the color matching scheme in the tool library; In response to receiving a model building instruction, the tool library retrieves the exterior component model to build a basic model of the mechanical exterior component according to the model building instruction, loads the brand element features into the basic model of the mechanical exterior component, and retrieves the color matching scheme from the tool library to render the basic model of the mechanical exterior component to obtain the mechanical exterior component application model.

[0006] Furthermore, the process of constructing the exterior component recognition model includes: Obtain exterior trim images from a product image database, perform standardization, normalization, and data encoding on the exterior trim images to obtain training standard images, and add real labels of exterior trim parts to the training standard images; Based on the neural convolution model, an exterior parts recognition model is constructed. The training standard images are used to train the exterior parts recognition model to obtain the shape training features and color training features of the exterior parts. The training loss value is calculated based on the shape training features, color training features and the true labels of the exterior parts. The expression formula is: ; In the formula, and It is expressed as setting weight; is the cross entropy loss function; F is the shape training feature; C is the color training feature; It is the shape feature in the real label of the exterior part; The color characteristics of the real label of the exterior parts; The Adam optimizer is used to optimize the parameters of the exterior parts recognition model according to the training loss value. The training process of the exterior parts recognition model is iterated repeatedly until the training loss value converges and the trained exterior parts recognition model is output.

[0007] Furthermore, the process of obtaining PEST category data includes: The trade restriction characteristics of mechanical exterior accessories are extracted from the policy trade data and used as the political factor data; the gross domestic product growth rate is used as the economic factor data of mechanical exterior accessories; the consumer evaluation data is used as the social factor data of mechanical exterior accessories, and the type characteristics of mechanical exterior accessories are obtained according to the start time of sale as the technical factor data; the PEST category data of mechanical exterior accessories is composed of political factor data, economic factor data, social factor data and technical factor data.

[0008] Further, the product standard image is read from the product image database according to the PEST category data, and the process includes: The product standard images in the product image database are configured with product description information and sales price; a first weight is set for the product standard image according to the economic factor data and the sales price, a second weight is set for the product standard image according to the social factor data, and the first weight and the second weight are superimposed to obtain a retrieval weight for the product standard image; The political factor data and the technical factor data are used as indexes to search the product description information to obtain the retrieved product standard image; the retrieved product standard image is screened according to the retrieval weight as the read product standard image.

[0009] Further, the product standard image is read from the product image database according to the product semantic data, and the process includes: The product semantic data includes product style intention data and product style intention vocabulary set; the product style intention data includes the shape description, form description, line description, color description and material description of the product design; the product style intention vocabulary set includes the design style description, functional characteristics description and user perception description of the product design; The product style intention data and the product style intention vocabulary set are classified and attribute annotations are added, wherein the attribute annotations include modern, simple, retro, light luxury and environmental protection; the words with the same attributes are sorted according to the repetition frequency, and a set number of words are screened out as indexes, and the product description information is searched to obtain the read product standard image.

[0010] Furthermore, the brand element characteristics include brand style intention data and brand value intention data. The brand style intention data includes the brand's logo, color scheme, image style and decorative elements; the brand value intention data includes font style and typesetting layout.

[0011] Furthermore, the exterior component model is generated according to the shape recognition features of the exterior component, and the process includes: The size features, contour features and material texture features of the exterior parts are calculated based on the shape recognition features of the exterior parts; a basic model of the exterior parts constructed in advance using a three-dimensional modeling platform is obtained; based on the size features and contour features mapped to the parameters of the basic model of the exterior parts, the material texture features are added to the surface of the basic model of the exterior parts to obtain the exterior parts model.

[0012] A second aspect of the present invention provides a mechanical exterior decoration model building system based on a neural network, comprising: An acquisition module acquires brand semantic data and PEST category data and product semantic data of mechanical exterior parts; reads product standard images from a product image database based on the PEST category data and product semantic data; extracts brand element features from a brand element database based on the brand semantic data; The feature extraction module inputs the product standard image into the preset exterior component recognition model to obtain the shape recognition features and color recognition features of the exterior component; generates the exterior component model according to the shape recognition features of the exterior component, and generates the color matching scheme according to the color recognition features; and stores the exterior component model and the color matching scheme in the tool library; The model building module, in response to receiving the model building instruction, retrieves the exterior component model from the tool library according to the model building instruction to build the basic model of the mechanical exterior parts, loads the brand element characteristics into the basic model of the mechanical exterior parts, and retrieves the color matching scheme from the tool library to render the basic model of the mechanical exterior parts to obtain the mechanical exterior parts application model.

[0013] Furthermore, it also includes a human-computer interaction interface, which is used to display the mechanical exterior parts application model and input model building instructions.

[0014] A third aspect of the present invention provides an electronic device, comprising a storage medium and a processor; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the mechanical exterior model building method described in the first aspect.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention obtains brand semantic data and PEST category data and product semantic data of mechanical exterior parts; reads product standard images from a product image database according to the PEST category data and product semantic data; extracts brand element features from a brand element database according to the brand semantic data; utilizes the PEST category data and product standard images to ensure that the design of exterior parts meets market positioning; and the brand element features enhance brand recognition.

[0016] The present invention inputs a product standard image into a preset exterior component recognition model to obtain shape recognition features and color recognition features of the exterior component; generates an exterior component model according to the shape recognition features of the exterior component, and generates a color matching scheme according to the color recognition features; stores the exterior component model and the color matching scheme in a tool library; can enrich the creative themes of designers, and assist designers in producing engineering machinery exterior components that meet brand semantics, product semantics, and market demand style types; stores the exterior component model and the color matching scheme in the tool library, so that these resources can be reused, thereby improving the flexibility and response speed of design.

[0017] In the present invention, according to the model building instruction, the tool library retrieves the exterior component model to build the basic model of the mechanical exterior component, loads the brand element characteristics into the basic model of the mechanical exterior component, and retrieves the color matching scheme from the tool library to render the basic model of the mechanical exterior component to obtain the mechanical exterior component application model; the automated data reading and model generation process reduces the need for manual operation, shortens the design cycle, and reduces costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of a method for constructing a mechanical exterior decoration model based on a neural network provided in Example 1 of the present invention; Figure 2 This is a structural diagram of a mechanical exterior decoration model building system based on a neural network provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0020] Example 1

[0021] like Figure 1 As shown, this implementation provides a method for constructing a mechanical exterior model based on a neural network, including: The process of building an exterior parts recognition model includes: Obtain exterior trim images from a product image database, perform standardization, normalization, and data encoding on the exterior trim images to obtain training standard images, and add real labels of exterior trim parts to the training standard images; Based on the neural convolution model, an exterior parts recognition model is constructed. The training standard images are used to train the exterior parts recognition model to obtain the shape training features and color training features of the exterior parts. The training loss value is calculated based on the shape training features, color training features and the true labels of the exterior parts. The expression formula is: ; In the formula, and It is expressed as setting weight; is the cross entropy loss function; F is the shape training feature; C is the color training feature; It is the shape feature in the real label of the exterior part; The color characteristics of the real label of the exterior parts; The Adam optimizer is used to optimize the parameters of the exterior parts recognition model according to the training loss value. The training process of the exterior parts recognition model is iterated repeatedly until the training loss value converges and the trained exterior parts recognition model is output.

[0022] The process of obtaining PEST category data includes: The trade restriction characteristics of mechanical exterior accessories are extracted from the policy trade data and used as the political factor data; the gross domestic product growth rate is used as the economic factor data of mechanical exterior accessories; the consumer evaluation data is used as the social factor data of mechanical exterior accessories, and the type characteristics of mechanical exterior accessories are obtained according to the start time of sale as the technical factor data; the PEST category data of mechanical exterior accessories is composed of political factor data, economic factor data, social factor data and technical factor data.

[0023] Read the product standard image from the product image database according to the PEST category data. The process includes: The product standard images in the product image database are configured with product description information and sales price; a first weight is set for the product standard image according to the economic factor data and the sales price, a second weight is set for the product standard image according to the social factor data, and the first weight and the second weight are superimposed to obtain a retrieval weight for the product standard image; The political factor data and the technical factor data are used as indexes to search the product description information to obtain the retrieved product standard image; the retrieved product standard image is screened according to the retrieval weight as the read product standard image.

[0024] By combining economic factor data and sales price to set weights, we ensure that the search results are highly correlated with market trends and price sensitivity. By setting search weights and indexes, we improve the efficiency and accuracy of retrieving relevant images from the product image database. By integrating political, economic, social and technological factors, we can more accurately locate market and consumer needs and improve the market adaptability of products.

[0025] Reading product standard images from the product image database according to product semantic data includes: The product semantic data includes product style intention data and product style intention vocabulary set; the product style intention data includes the shape description, form description, line description, color description and material description of the product design; the product style intention vocabulary set includes the design style description, functional characteristics description and user perception description of the product design; The product style intention data and the product style intention vocabulary set are classified and attribute annotations are added, wherein the attribute annotations include modern, simple, retro, light luxury and environmental protection; the words with the same attributes are sorted according to the repetition frequency, and a set number of words are screened out as indexes, and the product description information is searched to obtain the read product standard image.

[0026] Brand element features are extracted from a brand element database based on brand semantic data; the brand element features include brand style intention data and brand value intention data, the brand style intention data includes the brand's logo, color scheme, image style and decorative elements; the brand value intention data includes font style and typesetting layout.

[0027] Inputting the product standard image into the preset exterior component recognition model to obtain the shape recognition features and color recognition features of the exterior component; generating the exterior component model according to the shape recognition features of the exterior component, the process includes: The size features, contour features and material texture features of the exterior parts are calculated based on the shape recognition features of the exterior parts; a basic model of the exterior parts constructed in advance using a three-dimensional modeling platform is obtained; based on the size features and contour features mapped to the parameters of the basic model of the exterior parts, the material texture features are added to the surface of the basic model of the exterior parts to obtain the exterior parts model.

[0028] Generate color matching schemes based on color recognition features; store exterior component models and color matching schemes in the tool library; enrich designers' creative themes and assist designers in producing engineering machinery exterior components that meet brand semantics, product semantics, and market demand styles; store exterior component models and color matching schemes in the tool library so that these resources can be reused, improving design flexibility and response speed.

[0029] In response to receiving a model building instruction, the tool library retrieves the exterior component model to build a basic model of the mechanical exterior component according to the model building instruction, loads the brand element features into the basic model of the mechanical exterior component, and retrieves the color matching scheme from the tool library to render the basic model of the mechanical exterior component to obtain the mechanical exterior component application model.

[0030] The exterior trim contour information, exterior trim proportion information, exterior trim posture information and color identification information of the mechanical exterior trim application model are obtained; the exterior trim contour information includes body information, toolbox information, cockpit information and wheel information; the body information is divided into front face, rear face, configuration mechanism and headlights; the cockpit information includes air outlet, heat dissipation vent, support column, foot pedal and armrest; the wheel information includes wheel hub and wheel rim; the color identification information includes color, color value, painting and brand logo.

[0031] Then, the model is applied to the mechanical exterior parts to perform anomaly detection to obtain exterior detection data, and the exterior detection data is subjected to data analysis, which includes statistical analysis, correlation analysis and recognition analysis, wherein the statistical analysis method is used to analyze the distribution, mean and standard deviation of the data; the correlation analysis method is used to analyze the correlation between the data; and the recognition analysis method is used to identify patterns or anomalies in the data; The qualified range thresholds are set for key indicators in the exterior inspection data, and the data that exceeds the qualified range threshold in the exterior inspection data analysis data is marked as abnormal using the isolated forest detection method; the degree of abnormality of each abnormally marked data is confirmed based on the threshold value that exceeds the qualified range threshold; the abnormal data in different abnormality degrees are matched with attributes, and the attributes include design errors, material problems and process defects; finally, the exterior abnormality simulation data is obtained.

[0032] Example 2 like Figure 2 As shown, this embodiment provides a mechanical exterior model construction system based on a neural network, and the mechanical exterior model construction system is used to execute the mechanical exterior model construction method described in Example 1. The mechanical exterior model construction system includes: An acquisition module acquires brand semantic data and PEST category data and product semantic data of mechanical exterior parts; reads product standard images from a product image database based on the PEST category data and product semantic data; extracts brand element features from a brand element database based on the brand semantic data; The feature extraction module inputs the product standard image into the preset exterior component recognition model to obtain the shape recognition features and color recognition features of the exterior component; generates the exterior component model according to the shape recognition features of the exterior component, and generates the color matching scheme according to the color recognition features; and stores the exterior component model and the color matching scheme in the tool library; The model building module, in response to receiving the model building instruction, retrieves the exterior component model from the tool library according to the model building instruction to build the basic model of the mechanical exterior parts, loads the brand element characteristics into the basic model of the mechanical exterior parts, and retrieves the color matching scheme from the tool library to render the basic model of the mechanical exterior parts to obtain the mechanical exterior parts application model.

[0033] The human-computer interaction interface is used to display the application model of mechanical exterior parts and input model building instructions.

[0034] This embodiment can assist designers in producing engineering machinery exterior trims that meet brand semantics, product semantics, and market demand styles. The automated data reading and model generation process reduces the need for manual operations, shortens the design cycle, and reduces costs.

[0035] Example 3 This embodiment provides an electronic device, including a storage medium and a processor; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the mechanical exterior model construction method described in Example 1.

[0036] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, etc.) containing computer-usable program code. M , optical storage, etc.).

[0037] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0038] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0039] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0040] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for constructing a mechanical exterior model based on a neural network, characterized in that: include: Obtain brand semantic data as well as PEST category data and product semantic data of mechanical exterior parts; Read product standard images from the product image database based on PEST category data and product semantic data; extract brand element features from the brand element database based on brand semantic data; Input the product standard image into the preset exterior component recognition model to obtain the shape recognition features and color recognition features of the exterior component; generate the exterior component model according to the shape recognition features of the exterior component, and generate the color matching scheme according to the color recognition features; store the exterior component model and the color matching scheme in the tool library; In response to receiving a model building instruction, the tool library retrieves the exterior component model to build a basic model of the mechanical exterior component according to the model building instruction, loads the brand element features into the basic model of the mechanical exterior component, and retrieves the color matching scheme from the tool library to render the basic model of the mechanical exterior component to obtain the mechanical exterior component application model.

2. The method for constructing a mechanical exterior model according to claim 1, characterized in that: The construction process of the exterior component recognition model includes: Obtain exterior trim images from a product image database, perform standardization, normalization, and data encoding on the exterior trim images to obtain training standard images, and add real labels of exterior trim parts to the training standard images; Based on the neural convolution model, an exterior parts recognition model is constructed. The training standard images are used to train the exterior parts recognition model to obtain the shape training features and color training features of the exterior parts. The training loss value is calculated based on the shape training features, color training features and the true labels of the exterior parts. The expression formula is: ； In the formula, and It is expressed as setting weight; is the cross entropy loss function; F is the shape training feature; C is the color training feature; It is the shape feature in the real label of the exterior part; The color characteristics of the real label of the exterior parts; The Adam optimizer is used to optimize the parameters of the exterior parts recognition model according to the training loss value. The training process of the exterior parts recognition model is iterated repeatedly until the training loss value converges and the trained exterior parts recognition model is output.

3. The method for constructing a mechanical exterior model according to claim 1, characterized in that: The process of obtaining PEST category data includes: The trade restriction characteristics of mechanical exterior accessories are extracted from the policy trade data and used as the political factor data; the gross domestic product growth rate is used as the economic factor data of mechanical exterior accessories; the consumer evaluation data is used as the social factor data of mechanical exterior accessories, and the type characteristics of mechanical exterior accessories are obtained according to the start time of sale as the technical factor data; the PEST category data of mechanical exterior accessories is composed of political factor data, economic factor data, social factor data and technical factor data.

4. The method for constructing a mechanical exterior model according to claim 3, characterized in that: Read the product standard image from the product image database according to the PEST category data. The process includes: The product standard images in the product image database are configured with product description information and sales price; a first weight is set for the product standard image according to the economic factor data and the sales price, a second weight is set for the product standard image according to the social factor data, and the first weight and the second weight are superimposed to obtain a retrieval weight for the product standard image; The political factor data and the technical factor data are used as indexes to search the product description information to obtain the retrieved product standard image; the retrieved product standard image is screened according to the retrieval weight as the read product standard image.

5. The method for constructing a mechanical exterior model according to claim 1, characterized in that: Reading product standard images from the product image database according to product semantic data includes: The product semantic data includes product style intention data and product style intention vocabulary set; the product style intention data includes the shape description, form description, line description, color description and material description of the product design; the product style intention vocabulary set includes the design style description, functional characteristics description and user perception description of the product design; The product style intention data and the product style intention vocabulary set are classified and attribute annotations are added; the words with the same attribute are sorted according to the repetition frequency, and a set number of words are selected as indexes, and the product description information is searched to obtain the read product standard image.

6. The method for constructing a mechanical exterior model according to claim 1, characterized in that: The brand element characteristics include brand style intention data and brand value intention data. The brand style intention data includes the brand's logo, color scheme, image style and decorative elements; the brand value intention data includes font style and typesetting layout.

7. The method for constructing a mechanical exterior model according to claim 1, characterized in that: The exterior component model is generated according to the shape recognition features of the exterior component. The process includes: The size features, contour features and material texture features of the exterior parts are calculated based on the shape recognition features of the exterior parts; a basic model of the exterior parts constructed in advance using a three-dimensional modeling platform is obtained; based on the size features and contour features mapped to the parameters of the basic model of the exterior parts, the material texture features are added to the surface of the basic model of the exterior parts to obtain the exterior parts model.

8. A mechanical exterior model building system based on neural network, characterized in that: include: An acquisition module, which acquires brand semantic data, PEST category data of mechanical exterior parts, and product semantic data; Read product standard images from the product image database based on PEST category data and product semantic data; extract brand element features from the brand element database based on brand semantic data; The feature extraction module inputs the product standard image into the preset exterior component recognition model to obtain the shape recognition features and color recognition features of the exterior component; generates the exterior component model according to the shape recognition features of the exterior component, and generates the color matching scheme according to the color recognition features; and stores the exterior component model and the color matching scheme in the tool library; The model building module, in response to receiving the model building instruction, retrieves the exterior component model from the tool library according to the model building instruction to build the basic model of the mechanical exterior parts, loads the brand element characteristics into the basic model of the mechanical exterior parts, and retrieves the color matching scheme from the tool library to render the basic model of the mechanical exterior parts to obtain the mechanical exterior parts application model.

9. The mechanical exterior model building system according to claim 8, characterized in that , and also includes a human-computer interaction interface, which is used to display the mechanical exterior parts application model and input model building instructions.

10. An electronic device comprising a storage medium and a processor; the storage medium is used to store instructions; characterized in that: The processor is used to operate according to the instructions to execute the mechanical exterior model building method according to any one of claims 1 to 7.

Images