Research on 3D reconstruction method of human-computer interaction scene based on support vector machine in cloud manufacturing environment

Abstract

In three dimensional (3D) reconstruction of human-computer interaction scene images in cloud manufacturing environment, the traditional method recovers the depth information directly for projection mapping, which reduces 3D image reconstruction accuracy. A 3D reconstruction method for human-computer interaction scene images is proposed based on support vector machine, the human-computer interaction scene image in cloud manufacturing environment is pretreated, using extremum algorithm to detect noise in human-computer interaction scene images, through multi windows filtering method to filter pixels contaminated by noise. Using dynamic scene estimation method to enhance the human-computer interaction scene images in cloud manufacturing environment. Corner features of the human-computer interaction image are inquired, and extracted feature vectors are input into support vector machine to learn, and the immune algorithm is introduced to adjust the parameters and get the outlined structure of 3D human-computer interaction scene images, to realize the image 3D reconstruction. The experimental results show that in cloud manufacturing environment, the proposed method has high accuracy on reconstruction and authenticity.