Intelligent Design Method Of Automobile Modeling Style By Driven Individual Needs

The state issued the "Made in China 2025" strategic document focusing on intelligent manufacturing,proposing the concept of focusing on the development of intelligent equipment and intelligent products and promoting the intelligent production process.More and more car companies are continuously deep fusion of car manufacturing production technology and intelligent manufacturing technology.With the continuous improvement of domestic and foreign automobile companies' production and manufacturing level,the focus of competition in the automobile industry is shifting from manufacturing technology to automobile appearance design.Therefore,this article applies artificial intelligence technology to automobile styling design.The main tasks are as follows:(1)In order to solve the problem of TFIDF(Term Frequency-Inverse Document Frequency),the classic representative algorithm of unsupervised keyword extraction methods,which is completely based on word frequency information and ignores the impact of other lexical feature items on keywords,a fusion of multiple features is proposed.A method of extracting user requirements for automobile styling based on TFIDF algorithm.By acquiring unregistered vocabulary,introducing vocabulary features,and vocabulary emotional features,the accuracy of extracting user demand features is improved.The method is compared and analyzed with the traditional TFIDF algorithm and two existing improved algorithms to verify the effectiveness of the method.(2)In order to obtain car styling style images that meet the needs of users,this paper proposes a matching method for car styling features based on user needs,and constructs a multi-level labeling car styling library.The automobile modeling library mainly describes the characteristics of automobile modeling from two aspects: user attributes and automobile modeling style.The extracted user demand feature vector and the label vector of the automobile styling library are calculated for the similarity of the text space vector,and the image conforming to the automobile styling style is obtained,and the effectiveness of the method in this paper is verified by an example.(3)In order to solve the problem that the traditional image style transfer algorithm can only realize the feature transfer between two overall images,and thus cannot realize the transfer of the modeling features between the specified car objects,this paper proposes a car modeling image style based on semantic information Migration algorithm.The semantic segmentation network is introduced on the basis of the traditional style transfer algorithm,and the Markov random field is substituted for the Gram matrix of the traditional method to improve the style loss calculation method.This method is compared and analyzed with the traditional image style transfer algorithm and two existing improved algorithms to verify the effectiveness of the method.(4)In order to verify the effectiveness of the intelligent design method proposed in this paper,a case study of user demand feature extraction,automobile modeling feature matching,and automobile modeling intelligent generation is completed.And through the analysis of image quality,method performance and user evaluation,the effectiveness of the proposed method is verified.