Sea Mine Detection Based onMultiresolution Analysis and Noise
Sea Mine Detection Based onMultiresolution Analysis and Noise
复制标题
基于多分辨率分析和噪声的海雷探测
DOI：
--
复制
发表时间：
1999
期刊：
影响因子：
--
通讯作者：
Truong Q. NguyenyDept
中科院分区：
文献类型：
--
作者：
White Chen;Truong Q. NguyenyDept
我们的项目进一步扩展了Quyen，et.艾尔在水雷检测方面，采用匹配滤波技术对声纳图像中的水雷进行定位，并利用小波相关阈值去除噪声，提高了匹配滤波的性能。正在处理的数据是声纳图像集声纳0。在前人工作的基础上，我们进一步研究了多分辨率分析和噪声白化对基于匹配LTER的水雷探测算法性能的影响，提出了一种新的多尺度匹配LTER(MSMF)框架，并用相同的声纳图像数据对该算法进行了测试。由于算法的普及性、简单性和可视化程度的提高，我们在MatLab中进行了编程。虽然该算法的性能不如作者所期望的那样好，但该框架指出了未来的研究方向。
Our project further extends previous work 1] by Quyen, et. al. on sea mine detection, where matched ltering technique was employed to locate the mines in the sonar images, and wavelet coeecient thresholding was used to help removing the noise to enhance performance of the matched lter. The data being processed was the sonar image set Sonar 0. Based on the previous work, we further study the eeect of multiresolution analysis and noise whitening on the performance of such a matched lter-based algorithm for sea mine detection, and a new framework called \multiscale matched lter" (MSMF) is proposed. Again, exactly the same set of sonar image data is employed to test our algorithm. The programming was done in Matlab due to its increasing popularity, simplicity, and ease of visualization. Although the proposed algorithm has not yet performed well enough as the authors have been expecting, the framework indicates a direction for future research.