| With the improvement of mass consumption ability,personalized dress is a trend for people to show themselves.In order to highlight their personality and self-prominence,more and more people begin to choose customized service.The research is based on a set of 3D virtual clothing customization platform developed by the research group for Shandong Ruyi Group,which can realize the transformation of fabric,layout and virtual display.This topic is to increase the user's own virtual human head model and integrate it into the 3D virtual clothing customization platform to achieve a better user experience.On the premise of ensuring high scanning accuracy and speed,two Kinect second generation depth cameras are proposed to reconstruct three-dimensional human head model.At the same time,after virtual customization,when one key reservation volume service is performed,efficient and simple volume data recording means are needed in the field volume.At present,many high-level customization brands in the market mainly rely on traditional manual work to complete volume and data entry.Because of the factors such as human fatigue,input errors,long measurement time,customer fatigue and embarrassment,etc.,customization tasks often fail,so the traditional volume model is difficult to support the demand of mass customization services.In this paper,an intelligent AI voice volume system is proposed,which combines intelligent AI voice technology with traditional customized garment volume.The main research work of this paper is as follows:(1)Two Kinect second-generation depth cameras were used to reconstruct three-dimensional human head model:two Kinect were running at the same time,scanning the front and back parts of human head respectively and generating six three-dimensional depth point clouds.The adjustment algorithm was used to pre-process the depth point clouds after drying.Then the classical ICP algorithm was used to preliminarily register the front and back point clouds,and two depth point clouds were obtained:Font point clouds.Finally,an improved ICP algorithm is proposed to re-register Font point cloud and Back point cloud,i.e.point cloud registration fusion based on normal vector third-party reference and AABB bounding box collision algorithm,which makes the reconstruction of human head model more efficient and model quality better.(2)Develop a set of intelligent AI voice volume system:first,outline the business requirements of voice volume system;then analyze three main problems in the process of system development:inappropriate interface,data loading,scanning two-dimensional code;finally,focus on the design of intelligent AI voice volume system,analyze the system development environment,functional modules,database design in detail,and show the system.All part of the operation examples.According to the on-line use of voice volume system,the system greatly improves the accuracy and efficiency of volume data,thus indirectly speeding up the pace of intelligent development of enterprises,and further enhancing the economic benefits of enterprises.Through the field application of garment enterprises,this paper studies the reconstruction of three-dimensional human head model based on Kinect second-generation depth camera and the combination of intelligent AI voice and traditional garment customization volume,which meets the market demand and is suitable for popularization and use.In this paper,the development of new technology is applied to the traditional manufacturing industry,which has a certain degree of innovation.It is believed that the new technology will change the inefficient production mode of the traditional industry and provide endless power for the enterprise intellectualization under the impetus of the continuous in-depth three-dimensional reconstruction and intelligent AI voice technology research.There are 46 figures,21 tables and 66 references. |
PDF Full Text Request