面向环境仿真的目标特征重建方法研究
文献类型：学位论文
| 作者 | 岳田爽 |
| 学位类别 | 硕士 |
| 答辩日期 | 2014-05-28 |
| 授予单位 | 中国科学院沈阳自动化研究所 |
| 导师 | 赵怀慈 |
| 关键词 | 环境仿真 LiDAR 特征重建 |
| 其他题名 | Research of Target Feature Reconstruction Oriented to Environment Simulation |
| 学位专业 | 模式识别与智能系统 |
| 中文摘要 | 在自动目标识别领域，通常使用模板匹配方法完成对目标的识别与跟踪，模板的好坏在很大程度上决定了自动目标识别算法的效果，工程应用中，对于建筑物目标，一般使用轮廓线作为模板。本文借助环境仿真技术以及LiDAR点云数据处理技术，主要研究在考虑环境因素影响下对建筑物目标进行特征重建的方法，为搭建自动模板测试平台打下基础。 本文主要包括：基于LiDAR点云数据的建筑物轮廓线提取，环境仿真相关技术与方法研究，基于GPU的光线追踪算法研究与实现三部分内容。 基于LiDAR点云数据，本文实现了从LiDAR点云数据中提取建筑物轮廓线的技术流程，提出了建筑物轮廓线关键点提取方法，实验结果表明，该方法可以有效的滤除噪声，提取出建筑物轮廓线上的关键点。 针对在仿真中考虑光照等环境因素的需求，本文研究了环境场景仿真的相关技术，包括光照模型、光线追踪算法等，并提出了一种场景分割优化算法，实验结果表明，该方法可以有效提升光线追踪加速结构KD-Tree的创建速度。 为了提升仿真场景的渲染速度，本文完成了基于GPU的光线追踪系统的详细设计，使用GPU平台并行实现了光线追踪算法，实验结果表明，与CPU中的计算速度相比，在GPU中执行光线追踪算法可以获得数倍的性能提升。 |
| 索取号 | TP391.9/Y96/2014 |
| 英文摘要 | In the field of automatic target recognition, we usually completed target identification and tracking by using template matching method, the quality of the template largely determine the effect of automatic target recognition algorithm. In practice, we generally use contour lines as the template of building goals. This paper based on environmental simulation technology and LiDAR point cloud data processing technology , the main focus is researching building reconstruction methods in consideration of environmental factors, and build the foundation for automatic template test platform. The main content of this paper: building contour extraction based on LiDAR point cloud data, research of environment simulation related techniques and methods, research and implement of ray tracing algorithm based on GPU. Based on LiDAR point cloud data , the paper implemented the building countour line extraction process, and proposed a key points extraction algorithm of building’s contour line, experimental results show that this method can effectively filter out the noise and extract the key points of the building’s contour line. For the requirement of light and other environmental factors considered in the environment simulation, we studied the scene simulation environment-related technologies, including lighting model, ray tracing algorithm, and propose a scene segmentation optimization algorithm, experimental results show that this method can effectively improve the performance of KD-Tree construct process, which is the most famous ray tracing acceleration structure. In order to improve the speed of simulation scenario rendering process, the paper successfully implement the ray tracing algorithm with the help of GPU parallel implementation platform, meanwhile, the paper complete the detailed design of GPU-based ray tracing system, and finally implemented the algorithm, experimental results show that , compared with the CPU implementation, ray tracing algorithm get several times performance improvement in GPU. |
| 语种 | 中文 |
| 产权排序 | 1 |
| 页码 | 65页 |
| 分类号 | TP391.9 |
| 源URL | [http://ir.sia.ac.cn/handle/173321/14796]  |
| 专题 | 沈阳自动化研究所_光电信息技术研究室 |
| 推荐引用方式 GB/T 7714 | 岳田爽. 面向环境仿真的目标特征重建方法研究[D]. 中国科学院沈阳自动化研究所. 2014. |
入库方式： OAI收割
来源：沈阳自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。