基于增强现实和神经网络的校园展示应用设计

Design of Campus Display Application Based on Augmented Reality and Neural Network

当前的校园展示多采用虚拟现实开发,该技术让用户完全沉浸入虚拟世界,但忽略了现实世界的信息。增强现实技术能够在真实世界的基础上叠加虚拟信息,进一步增强人们的感知体验。基于Unity3D+ARCore+Tensor Flow设计了带平面检测、多图像和物体识别功能的增强现实应用,详细描述了虚拟对象、触屏交互、平面检测、图片和物体识别的实现过程和注意问题,说明了在脚本和UI设计中的性能优化方法。测试表明多标识物识别提升了应用的实用性,物体识别功能丰富了应用内涵,增加了校园展示的设计方法,拓展了增强现实技术的应用范围。

Currently most of the campus displays are developed through virtual reality technology,which gets users immersed in the virtual world,but ignore information of the real world.In this paper,a campus display application is designed based on Unity3 D+ARCore+TensorFlow,which provide functions of plane detection,multi-image and object recognition.The implementation process and problems of virtual objects,touch screen interaction,plane detection,image and object recognition are described in detail.Performance optimization methods in scripting and UI design are illustrated.The test shows that multi-marker recognition improves practicability,and object recognition enriches application implications,which together could increase the design methods of campus displays and expand the application scope of augmented reality technology.