VR激光成像噪声过滤技术研究  

Design of Multi-Laser Sensor Controller for Smart Home

作　　者：苏叶健 张叶茂心[1,3] SU Yejian;ZHANG Yemao(Guangxi Vocational Education Research School For Artificial Intelligence,Nanning 530008;Nanning College For Vocational Technology,Nanning 530008,China;University of Electronic Science and Technology of China,Chengdou 610054,China)

机构地区：[1]广西人工智能职教育研究院,南宁530008 [2]南宁职业技术学院,南宁530008 [3]电子科技大学,成都610054

出　　处：《激光杂志》2020年第2期190-193,共4页Laser Journal

基　　金：2019年度广西高校中青年教师科研基础能力提升项目(No.2019KY1249)

摘　　要：VR激光成像中存在大量随机噪声,为了提高VR激光虚拟成像能力,提出一种VR激光成像噪声过滤方法,对采集的三维小区域物体的特征图像,分块匹配和自适应信息融合,完成虚拟VR激光图像重构处理,结合模板匹配技术,对VR激光虚拟成像的特征匹配和滤波处理,采用多层Gabor小波降噪方法进行激光小区域成像后的图像降噪处理,对降噪滤波后的激光小区域成像输出采用改进的虚拟现实仿真技术实现图像重建,实现噪声过滤。仿真结果表明,该方法去噪分辨率更强,抗干扰性与传统方法对比,迭代60次后,成像信噪比达到最大49.5 dB,更优于贝叶斯方法的25.4 dB与主成分分析方法的25.4 d B。负载开销与时延的对比也较其他两种方法更为经济。There are a lot of random noises in VR laser imaging.In order to improve the virtual imaging ability of VR laser,a noise filtering method of VR laser imaging is proposed.The feature images of three-dimensional small area objects are collected,matched in blocks and fused with adaptive information.The reconstruction of virtual VR laser image is completed.Combining with template matching technology,the feature matching and filtering of VR laser virtual imaging are processed.The multi-layer Gabor wavelet denoising method is used to denoise the image after laser small area imaging.The improved virtual reality simulation technology is used to reconstruct the image after denoising and filtering the output of laser small area imaging.The simulation results show that,After 60 iterations,the imaging signal-to-noise ratio reaches a maximum of 49.5 d B,which is better than 25.4 d B of Bayesian method and 25.4 dB of principal component analysis method.The comparison of load overhead and time delay is also more economical than the other two methods.

关 键 词：VR 激光 小区域成像 小波降噪 噪声过滤 

分 类 号：TN247[电子电信—物理电子学]