Research On 3D Human Body Model Reconstruction System Based On Sequence Image

With the higher demands of people for products,they are favoring toward personalized custom products and 3D portrait museum is entering into people's life.The printing quality of 3D portrait printing is affected by the accuracy of the 3D human body model and the accuracy of the 3D printer.The recognition of the 3D human body model depends on the quality of the modeling data and the data processing method.Because the three-dimensional scanners in the market are not only expensive but also complex in use,3D portrait reconstruction process exists high-cost,time-consuming and other issues.In view of existing shortcomings,this paper develops a low cost and high precision 3D model reconstruction system.It is based on sequence image reconstruction of three-dimensional model of the human body.This paper designs a set of hardware systems for sequence image acquisition based on the full theoretical study of image-based 3D reconstruction.The system can be used to acquire high-quality,high-overlap image sequences with multi-angle modeling objectives.Then use the sequence image for three-dimensional reconstruction.First,we pre-process the sequence image,convert the original raw format image into Tiff format image,eliminate the interference of image noise,and remove the redundant background to reduce unnecessary time of the model project reconstruction process.Next,the sparse point cloud data of the modeling target is obtained by using the algorithm based on independent components analysis(ICA)and the multi-view 3D reconstruction,and the accuracy of the model is improved by the iterative algorithm.Then,in order to get the dense point cloud,the PMVS algorithm is used to densify the obtained point cloud data.Then,the obtained point cloud is used to reconstruct a 3D human body model by using the Poisson algorithm.Finally,this paper uses Geomagic software to refine and transform the three-dimensional model of reconstruction to obtain a three-dimensional human body model with high degree of recognition for 3D printing.Based on sequence image reconstruction of the three-dimensional model of the whole body,this paper combines the three-dimensional point cloud data reconstructed with the human head as the target object and the point cloud data obtained from the human body as the target object.Finally,the body three-dimensional model with clear facial features and high degree of identification is obtained after the Poisson surface reconstruction.In this paper,we find the final three-dimensional model reconstructed by three-dimensional model reconstruction system based on the sequence image through 3D printing works well by printing verification.The result of this test shows that the system can obtain high-definition three-dimensional human body model to meet the needs of 3D portrait printing hall,and it is worth promoting the use.