Research On Non-contacting Human Body Measurement And Modeling Technology Based On Image

The non-contacting body measurement technology gradually replaces the traditional contacting body measurement method with its advantages of rapidness,accuracy,and high efficiency.It is widely used in the establishment of human body databases,human body modeling,and virtual fitting and other fields.However,due to the high cost of equipment and strict environmental requirements,it is difficult to carry out human body measurements in large quantities.In order to solve these problems,this thesis proposes a method of anthropometry and modeling based on human body front and side images.The main research work includes the following aspects:(1)Aiming at the problem that the traditional human contour extraction method has strict requirements on the environment,this thesis proposes a method for extracting human contours under complex background environment,which is mainly to separate the human body contour into different parts.Because the H component of the HSV color space is independent of light,hair and shorts contours can be extracted.Skin contours are extracted based on the elliptical skin detection model and the HSV threshold model.Finally,the coat contour is extracted based on the HSV threshold model.Experiments show that the method proposed in this thesis can reduce the conditional constraints of the background environment and the human body contour segmentation results are good.(2)Aiming at the problem of human body feature point misdetection and missed detection by traditional methods,this thesis proposes a human feature point extraction method based on human body proportion relationship and Freeman chain code.Firstly,the human body contour is represented by the Freeman chain code.The location of the inductive feature points is summarized according to the change rule of the chain code values,and then the human body proportion relationship is used to eliminate the interference points and supplement the undetected feature points.Then the length parameters and girth parameters are calculated through the feature points.The measurement results are displayed through a visual interface.Experiments show that the measurement error of the proposed method is basically controlled at about 5%,which meets the requirements of human body measurements.(3)A human body modeling method based on human body contour and Mayamodeling software is proposed.Based on the human body contour and the characteristics of the human body to approximate the ellipse,the cross-sectional data of the front profile is used as the long axis of the ellipse and the cross-sectional data of the side profile is as the minor axis of the ellipse,a three-dimensional digital model of the human body is generated.The human body digital model was imported into the Maya software through the MEL script language,and the human body model was stitched and texture-mapped to obtain a three-dimensional human body model.Experiments show that the proposed method improves the efficiency of modeling and the model distortion is low.