【作者】 刘君；

【导师】 高林；

【作者基本信息】 北京林业大学 ， 机械电子工程， 2012， 硕士

【摘要】 本文通过研究探地雷达对浅地层物体的图像数据处理,形成的三维立体图,逐步转到林业领域,为将雷达应用于林业领域的打下基础。对于探地雷达探测,地下环境比较复杂,反射波形成的结果也比较复杂,一般都需要对波形进行处理,使用滤波器对杂波进行抑制处理后,才能对雷达波数据进行进一步的处理。对雷达波数据进行量化处理后,提取数据,并通过识别算法识别出有用信号,然后转化为空间坐标数据,进一步生成三维图像,并在之后对三维图进行进一步的改进等,转换成连续的立体图。本文研究了不同介质对雷达波的影响情况,提出了基于k-means算法的数据提取算法,并对该算法作了进一步改进。然后使用Matlab编程对目标的三维成像进行成像,并对图像进行进一步的优化和完善,取得了较好效果。通过对雷达在地质方面的应用情况,证实了三维立体成像方法的可行性,为进一步雷达在林业方面的使用打下基础。本文独创性在于：1.改进了k-means算法,提出优化算法,提升识别率,对数据的提取更准确；2.根据雷达探测对不同介质形成图像不同的特点,写出了针对不同物体介质的识别算法。更多还原

【Abstract】 By studying ground-penetrating radar images of shallow strata of objects data processing and formation of three-dimensional image, slowly turn to the field of forestry, supply foundation for forestry. Underground environment is more complex, reflected wave signal is complicated, the general waveform processing, use filters noise suppression processing later. Generate3D images, and after further improvements, render three-dimensional image into a continuous one. Studies impact of different media on radar, K-means algorithm and the improved, use Matlab programming, image for further optimization and improvement, has achieved good results. Supply the foundation for further radar used in forestry field through application of radar in geological, confirming the feasibility of three-dimensional imaging method. This research has a few innovations as below:1. Improve k-means algorithm, propose a new algorithm which make a higher recognition rate on the extraction of data more accurate;2. Radar image different characteristics of different media, writing recognition algorithm for different objects in the media.更多还原