QQ客服 官方微博 反馈留言

题名	基于侧扫声图的图像处理方法的研究
作者	王斯朕
学位类别	博士
答辩日期	2009-05-24
授予单位	中国科学院声学研究所
授予地点	声学研究所
关键词	侧扫声纳 声纳图像 图像处理 伪彩色变换 图像增强 图像降噪 多尺度形态学运算 图像分割 分形维数 目标分类
其他题名	Research on Methods of Image Processing Based on Side-scan Sonar Image
学位专业	信号与信息处理
中文摘要	声纳设备通常利用声波对海底地形及水下目标图像进行获取。侧扫声纳是一种应用非常广泛的声纳设备。对侧扫声纳图像处理方法的研究是声纳研究领域一个有意义的课题。本文介绍和提出了一系列针对侧扫声纳图像的图像处理方法，并采用这些方法对实际侧扫声图进行处理。得到的实验结果和结论是今后关于侧扫声图进一步研究的理论和实验基础。本文的主要内容包括： 1． 水声学、侧扫声纳以及水声图像处理的历史和技术概述。 2． 在侧扫声图的图像显示方面，介绍侧扫声纳的声图数据格式以及声图的灰度变换方法。提出了基于B样条插值曲线的图像伪彩色变换方法。 3． 采用直方图均衡化、灰度畸变均衡、阴影加强等方法对图像增强。在图像降噪部分，采用均值滤波、顺序统计滤波和高斯滤波等方法去除图像噪声，并对各降噪方法的性能进行测试。 4． 利用数学形态学运算方法对侧扫声图进行处理。采用多尺度形态学运算对图像的细节特征进行操作。 5． 分别采用传统边缘检测方法和形态学边缘检测方法对目标边缘进行检测。选取迭代法、Ostu法等阈值计算方法分割图像。采用基于分形维数的目标区域探测方法探测侧扫声图中的目标区域。 6． 根据不变矩的理论，对侧扫声图中的目标进行分类识别。并采用基于直方图统计特征的目标分类方法对目标分类。
英文摘要	Submarine topography and underwater target image is always acquired by sonar apparatus. Side-scan sonar is one type of sonar which is widely used. It is significant to do research on the method of processing of side-scan sonar image. A series of image processing method based on side-scan sonar image are introduced and proposed in this dissertation. Some actual sonar images are processed by these methods, and the results are given in the paper too. The processing results lay a theoretical foundation and an experimental foundation for the further research of sonar image processing. The major contents of the dissertation are as follows: 1. The dissertation summarized the history and technology of hydroacoustics, side-scan sonar and processing of the sonar image. 2. The data format of the sonar image and the methods of gray transform are introduced in this paper. A pseudocolor transformation method based on B-spline curves is proposed too. 3. The methods of histogram equalization, graylevel distortion equalization and shadow enhancement are used in the enhancement of sonar image. Several methods are used to reduce the noise in sonar image, such as average filtering, order statistic filtering, and Gaussian low-pass filtering. The performance of the noise filters is tested. 4. Sonar images are processed by the methods of mathematical morphology. Methods of image processing based on multi-scale mathematical morphology are proposed to process the details and the features of the sonar image. 5. The edge of underwater target is detected by traditional edge detection method and the edge detection method based on mathematical morphology. Several methods are used to split sonar image, including the method of iteration, the Ostu method, and the generalization of the Ostu method. The target area in sonar image is detected by the target area detection method based on fractal dimension. 6. Underwater targets in sonar image are classified, based on the theory of image moment invariant. A target classification method based on histogram statistical characteristic is used to classify the underwater targets too.
语种	中文
公开日期	2011-05-07
页码	105
内容类型	学位论文
源URL	[http://159.226.59.140/handle/311008/566]
专题	声学研究所_声学所博硕士学位论文_1981-2009博硕士学位论文
推荐引用方式 GB/T 7714	王斯朕. 基于侧扫声图的图像处理方法的研究[D]. 声学研究所. 中国科学院声学研究所. 2009.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

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。