Virtual Assembly Technology Based On Skeleton Tracking

As China’s manufacturing 2025 project is put forward, higher requirements for industrial manufacturing technology are advocated. Assembly link has occupied a vital position in the process of product development, especially for large-scale scene assembly. If there appear any errors, it will lead to the waste of manpower and material resources. Responding to this problem, industry field has developed various types of respective virtual assembly systems. However, ordinary virtual assembly system is not only expensive but also has high threshold for learning. In order to solve some problems of virtual assembly, this article mainly conducts its research from the following four aspects.Firstly, it will give an analysis about the main composition and the structure for the assembly objects with the 3D modeling. This article selects turbofan aeroengine as assembly object. Turbofan aeroengine is composed of different units, it uses 3D modeling software to realize 3D modeling of the main unit body according to the analysis results of the whole structure of the turbofan engine. Besides, it gives hierarchical classifications for the 3 d model. It completes the entire assembly process for the optimize of assembly procedure with the combination of actual assembly requirements and characteristics of the virtual assembly system.Secondly, this research has analyzed the data acquisition of human body skeleton based on Kinect. Traditional sports data capture device have been unable to meet the needs of contemporary virtual assembly. There needs to seek a new way of motion data acquisition, and it has a natural advantage to use Kinect as motion data acquisition platform for virtual assembly. Based on the Light Coding technology of obtaining the basic image data and 3 d depth image, this platform separates human body image from the depth image so as to build skeletal tracking system.Thirdly, it focus on the design of the assembly system interaction. According to the requirements of modern virtual assembly and the characteristics of system itself, this paper has designed a set of limb movement interaction scheme based on skeleton tracking rather than simply gesture recognition. First, it acquires body depth image through the depth camera, and then it gains human skeletal system by the processing of depth image. The interaction of assembly system is realized by binding the acquired human bones with virtual human body skeleton under virtual environment, adding the use of collision detection in virtual engine response and sparse matching point.Fourthly, the complete construction of virtual assembly training system based on skeleton tracking. This article has analyzed platform construction of the upper machine and lower machine to build virtual assembly environment. Finally, this article verifies the feasibility and the effectiveness of the proposed method through the concrete experiments.