复杂场景中多椭圆目标的检测

复杂场景中多椭圆目标的检测(任务书,开题报告,外文翻译,论文12700字)
摘要
椭圆检测是图像处理的一个热点问题，由于椭圆目标广泛存在于社会、经济、军事等各个领域，且人们出于各种需求对它进行提取，因此如何高效、准确的检测出这些椭圆形物体，具有重要的应用及学术研究价值。
本文基于主元素法，对前景提取后的样本区域进行多椭圆拟合。首先，在样本提取阶段，采用阈值法从场景中提取出二值前景目标；其后，采用一个椭圆来拟合场景中的目标区域，通过设计的拟合评价指标来判断拟合的误差，如果误差小于预定阈值，则结束运行；如果误差大于阈值，则根据拟合误差新添一个椭圆，并基于EM算法，优化各个椭圆的参数，包含椭圆中心，长轴、短轴、及长轴方向；再次进行拟合误差计算，并重复上面各步骤，直至拟合误差达到要求。
采用本文设计的方法，对各种复杂的前景目标进行椭圆分析，可以较好地从中检测、分割出各种类椭圆目标。其肉眼精度及量化误差都符合实际应用的要求，相关实验验证了本文算法的有效性。
关键词：多椭圆检测EM算法主元素法
 
Abstract
Ellipse detection is a hot issue in image processing. Because elliptical targets are widely used in various fields, such as society, economy, military and other fields, people extract them out of various needs. Therefore, how to detect these elliptical objects efficiently and accurately has important and academic research value. [来源：http://Doc163.com]
Based on the principal element method, multi ellipse fitting is applied to the sample area after foreground extraction. First, in the sample extraction stage, the two value target is extracted from the scene by the threshold method. Then, an ellipse is used to fit the target area in the scene, and the fitting error is judged by the designed fitting evaluation index. If the error is less than the predetermined threshold, the error is over. If the error is greater than the threshold, then the root is the root. According to the fitting error, a new ellipse is added, and the parameters of each ellipse are optimized based on the EM algorithm, including the ellipse center, long axis, short axis and long axis direction, and the fitting error is calculated again, and the above steps are repeated until the fitting error is reached.
The elliptical analysis of various complex foreground targets can be achieved by using the method designed in this paper. Its gross accuracy and quantization error meet the requirements of practical application. Relevant experiments verify the effectiveness of the algorithm. [资料来源：http://Doc163.com]
Key words: multi-ellipse detection, EM algorithm, principal element method
  [资料来源：http://www.doc163.com]

目录
摘要    III
第1章 绪论    1
1.1课题研究的背景及意义    1
1.2椭圆检测算法的研究现状    1
1.3算法内容及论文结构安排    2
第2章 椭圆拟合算法简介    4
2.1图像预处理    4
2.1.1灰度化图像    4
2.1.2二值图像    4
2.1.3边缘检测    5
2.2Hough变换方法    6
2.2.1Hough变换算法描述    6
2.2.2Hough变换算法存在的问题    8 [来源：http://Doc163.com]
2.3最小二乘法    8
2.3.1最小二乘法描述    8
2.3.2最小二乘法存在的问题    9
2.4本章小结    9
第3章 EM算法原理    10
3.1   EM算法的数学理解    10
3.1.1  最大后验概率    10
3.1.2   Jensen不等式    11
3.2   EM算法的思想及算法描述    12
第4章 基于EM算法及主元素法的椭圆拟合法    14
4.1主元素法拟合椭圆原理    14
4.2多椭圆目标检测的EM算法设计    16
第5章  本文算法的实际拟合结果与总结    18
5.1   单椭圆目标拟合结果    18
5.2   多椭圆目标拟合结果    20
5.3   拟合结果结论    21
第6章 总结与展望    22
6.1全文总结    22
6.2不足与展望    22
参考文献    23

[资料来源：www.doc163.com]