Font Size: a A A

The Study Of Material Texture Image Model By Kansei Engineering

Posted on:2014-01-20Degree:MasterType:Thesis
Country:ChinaCandidate:R Y TangFull Text:PDF
GTID:2251330422450914Subject:Mechanical design and theory
Abstract/Summary:PDF Full Text Request
The emergence and development of Kansei engineering provide new ideas for theresearch of industrial design. By means of Kansei Engineering, the vague uncertainemotional problems can be quantified and studied by engineering research methods.Today, Kansei engineering has been increasingly used in the field of research ofmaterial texture image. This topic is based on the common means of Kansei engineering,taken common plastic material as object, established a mathematical model betweenobjective material parameters and subjective feelings of meterials, and discussed therelationship between them.In the full study, based on the idea of orthogonal test and the methods of computerrendering, obtain material samples. Using statistical methods to elected samples whichare representative scientifically; Collection a large numbers of vocabularis which areused to describe the feeling of plastic frequently and select the representative ones asmeasures of subjective feelings of material. Taken representative samples as subjects,in collection with the representative emotional words, quantify the materials feelingcharacteristics, obtain corresponding data of objective parameters and subjectiveevaluation of materials.Based on the data of material texture experiment and the methods of BP neuralnetwork, establish mathematical model between objective parameters and subjectiveevaluation of materials.For the actual situation, select suitable network structure,activation function and network learning methods, find a material texture image modelwith good fitting performance by the training of the network. In this way, thequantitative of the relationship of objective paraments and subjective evaluation ofmaterials can be achievement.Based on good global optimization capability and good macro-search capability ofgenetic algorithm, optimization the thresholds and weights of BP material texturemodels。The acuracy and pan capabilities of network can be improved and thequantitative description of relationship between objective paraments and subjectiveevaluation of materials can be more accurate and effective. By the verificationexperiment, the accuracy and reliability of the model can be guaranteed.By analyzing the result of GA-BP Model, the mutual effect rule between objective parameters and subjective evaluation measure can be figured out. To summary the effectto some particular textural image of different materials in different conditions, afoundation which can be used to predict the character of some already know materialcan be found.And this can be used as a law of choosing material during the designingprocess will also be identified.
Keywords/Search Tags:Kansei engineering, material texture image, BP neural network, geneticalgorithm
PDF Full Text Request
Related items