Hough变换关键问题研究及其在航迹起始方法中的应用

【作者】 金术玲；

【导师】 程咏梅； 潘泉；

【作者基本信息】 西北工业大学 ， 控制理论与控制工程， 2004， 硕士

【摘要】 通过大量数据的非相干积累，利用Hough变换可在低信噪比下提取信号特征，这在低信噪比自动航迹起始以及低信噪比图像处理中有重要的应用价值。本文在“十五”预研项目关于低信噪比目标跟踪研究的资助下，研究了标准Hough变换的参数空间量化以及参数提取问题。主要工作如下： 1．以基于Hough变换的航迹起始为具体研究对象，比较深入地分析了标准Hough变换在应用中存在的三个关键问题：参数量化、积累方式和参数提取。 2．推导了参数积累单元投票的概率分布，并根据所得的分布指出了因变量参数量化时应注意的问题，为航迹可分辨性的参数设计提供了依据。 3．提出了一种参数积累单元的权值积累方法，改善了峰值单元与其附近单元的积累值的可区分性，从而为参数提取创造了有利条件。仿真表明该方法能够有效起始航迹。 4．给出了一种峰值提取算法，并在航迹起始中，将采用该算法与不采用该算法的航迹起始结果进行了比较，验证了该算法的可行性和有效性。更多还原

【Abstract】 Through the noncoherent integration of great numbers of data, and by using Hough transform the features of signals in low SNR (signal-to-noise-ratio) can be extracted. Therefore, the Hough transform may play an important role in the automatic track initiation and the processing of images with low SNR environment. Supported by the key research project related to target tracking in low SNR and electronic warfare’s environment, a part of the tenth Five-year plan our research focuses on the problems of the parameter quantization and parameter extraction. The main works are as follows.1. Three key problems of Hough transform in track initialization field are presented: parameter quantization , accumulate method and parameter extraction .2. Firstly, the vote distribution of parameter accumulate cells is proposed. Hence several important aspects of the dependent variable are proposed and which gives a reference for how to quantize the parameters.3. Secondly, a method of weight integration of ’votes’ is proposed. Simulation results show that the method can significantly improve the separability between the local peak cell and its neighbors compared with the standard binary integration method.4. Finally, a peak extraction method is proposed for further improving the separability between the local peak cell and its neighbors in multiple target tracking. The simulation of multiple track initiation shows its feasibility and effectiveness.更多还原