基于激光成像及VR技术的多视点图像场景重构技术

Multi-view image reconstruction based on laser imaging and VR

为了提高多视点激光图像的场景重构能力,提出基于激光成像及VR技术的多视点图像场景重构技术。建立多视点激光场景图像虚拟现实三维重构模型,在大气散射环境下进行多视点激光场景图像的光强自适应融合,采用分块特征匹配技术进行多视点激光场景图像的信息增强处理,采用激光散射光晕点匹配方法进行多视点激光场景图像的细化滤波处理,采用激光散射光晕点特征检测方法进行图像特征提取,使用亮度分量进行多视点激光场景图像特征细节透射分析,对提取的多视点图像场景特征信息采用卷积神经网络学习方法进行多视点激光场景图像场景融合和特征重构,实现多视点激光场景图像虚拟现实三维重构。仿真结果表明,采用该方法进行多视点激光场景图像虚拟现实三维重构的精度较高,峰值信噪比较高,图像细节特征分辨力和准确性较好。

In order to improve the scene reconstruction capability of multi-view laser images,a multi-view image scene reconstruction technology based on laser imaging and VR technology is proposed.Establishing a virtual reality three-dimensional reconstruction model of multi-view laser scene image,carrying out light intensity adaptive fusion of the multi-view laser scene image in an atmospheric scattering environment,carrying out information enhancement processing of multi-view laser scene image by adopting the block feature matching technology,carrying out thinning filtering processing of the multi-view laser scene image by adopting laser scattering halo point matching method,carrying out image feature extraction by adopting laser scattering halo point feature detection method,and the brightness component is used for transmission analysis of feature details of multi-view laser scene images.Convolution neural network learning method is used for scene fusion and feature reconstruction of extracted multi-view laser scene image feature information to realize 3 D reconstruction of multi-view laser scene image virtual reality.The simulation results show that the 3 D reconstruction of multi-view laser scene images using this method has high accuracy,the PSNR is higher and the resolution and accuracy of image detail features are good.