基于高低频特征增强算法的室内场景三维重建仿真
Simulation of Indoor Scene 3D Reconstruction Based on High and Low Frequency Feature Enhancement Algorithm
郑晓倩1
作者信息
- 1. 福建水利电力职业技术学院 建筑工程学院,福建 永安 366000
- 折叠
摘要
为了提高室内设计效率和质量,提出了一种基于高低频特征增强算法的室内场景三维重建仿真方法.利用小波变换算法提取室内场景单图像高低频特征,通过加权引导滤波算法增强高低频特征图像;采用深度神经网络,预测图像的分割图和深度图,并结合长短时记忆网络估计室内场景单图像平面参数.然后整合分割图、深度图与平面参数,实现三维重建.实验结果表明:该方法可有效提取室内场景单图像高低频特征,增强高低频特征图像和提升图像清晰度;三维重建均方归一化误差最高值为 0.15.
Abstract
In order to improve the efficiency and quality of interior design,a 3D reconstruction simulation method for indoor scenes based on high and low frequency feature enhancement algorithm is proposed.Combined with long short-term memory networks to estimate single image plane parameters of indoor scenes,the paper extracts high and low frequency features of a single indoor scene image using wavelet transform algorithm,enhances the high and low frequency feature images through weighted guided filtering algorithm and uses deep neural networks to predict segmentation and depth maps of images.Afterwards,integrate the segmentation map,depth map,and plane parameters to achieve 3D reconstruction.The experimental results show that this method can effectively extract high and low frequency features of indoor scene single images,enhance high and low frequency feature images,and improve image clarity.The highest mean square normalization error of 3D reconstruction is 0.15.
关键词
高低频特征增强算法/室内场景/单图像/分段平面/三维重建Key words
high and low frequency feature enhancement algorithm/indoor scenes/single image/segmented plane/three-dimensional reconstruction引用本文复制引用
出版年
2024
NETL
NSTL
万方数据