节点文献

激光雷达距离像去噪和边缘检测算法研究

Reserch on Denoising and Edge Detection of Range Image LIDAR

【作者】 邓礼茂

【导师】 鲁新平;

【作者基本信息】 国防科学技术大学 , 电子与通信工程, 2011, 硕士

【摘要】 成像激光雷达是激光技术与雷达技术相结合而产生的新型雷达。成像激光雷达自动目标识别就是对成像激光雷达所获取的原始图像进行图像预处理、图像分割、特征提取和模式分类等处理,最终实现对目标的自动识别。图像分割是图像处理和分析的关键步骤,对后续高层次的特征描述、匹配和识别等有着重大的影响。本论文主要内容有:1、分析了成像激光雷达图像自动目标识别的研究背景与意义,研究了成像激光雷达成像机理、成像激光雷达距离选通、噪声抑制、图像分割的国内外研究现状和技术途径。2、在分析典型成像激光雷达系统的基础上,将用于抑制后向散射的距离选通技术应用到了三维成像激光雷达上,并采用二值化算法合成距离三维像。3、提出了一种基于Contourlet变换的成像激光雷达自适应阈值去噪方法。实验表明,该算法效果较好,能有效地抑制距离反常噪声,较好地保护激光图像中目标的边缘细节。4、通过对激光雷达图像的成像特性分析,提出了一种基于改进拉普拉斯金字塔( LP )塔形分解的多尺度融合的边缘提取算法,该算法能有效地检测边缘。

【Abstract】 The imaging laser radar is a new kind of radar which combined laser technique and radar technique. The automatic recognition of the imaging laser radar used the technology of image segmentation, feature extraction and pattern list to deal with the image which has been captured by it to realize the automatic target recognition. Image segmentation is a very important step in image process. It has great effect to the feature extraction, marching and recognition.The main work of this paper is:1、Introduce the background of the research on imaging laser radar to recognition target automatically and describe the system of the imaging laser radar. Summary the approach of the range gated, noise reduction and segmentation in the imaging laser radar at home and abroad.2、After analyzing the system of the typical imaging laser radar, use the technology of noise reduction range gated to the 3-D imaging laser radar and use the algorithm of binarizing to the distance 3-D image.3、This paper proposed a new method to reduce the noise by contourlet transform. By experiment, it shows that the new algorithm can achieve better effect than traditional ways and protect the target’s edge minutiae in the image.4、After analyzing the imaging characteristics of laser radar image, a multiscale edge detection algorithm based on Laplacian Pyramid was proposed. This algorithm can detect edges of images reliably and effectively.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络