摘要遥感图像中的目标检测一直是模式识别和遥感图像处理的热点课题,而舰船目标无 论是作为运输载体还是作为军事目标,其自动检测在管理海域交通,救援遇难船只,以及 海上作战等领域都有着非常广阔的研究和应用前景。78632
本文对红外遥感图像中的舰船检测方法进行了探究和改进。主要采用融合区域生长 和形态学图像处理的方法来进行海陆分割并判断图像是海陆共存还是仅有海洋区域。如 果仅有海洋区域,则先对其进行高斯低通滤波处理来减弱条带噪声的影响,再利用对连 通区域的处理进行舰船检测;如果是海陆共存,则分别对海岸线附近区域和海洋区域进 行连通区域的处理以检测舰船,最后将两部分检测到的结果合并得到最终舰船检测的结 果。
毕业论文关键词 红外遥感图像 舰船检测 区域生长 形态学 连通区域 条带噪声
毕 业 设 计 说 明 书 外 文 摘 要
Title Research on Ship Detection Methods from Infrared Remote Sensing Image
Abstract Target detection from remote sensing images is always the hot topic in the fields of pattern recognition and remote sensing image processing。 Whether the ships are as transport carriers or military targets, automatic detection of ships has an important and wide applications in areas such as management of vessel traffic service, rescue of the shipwreck, and naval warfare。
In this paper, ship detection from infrared remote sensing image is explored and improved。 A method combing the region growing and morphological image processing algorithm is mainly used in this paper to do sea-land segmentation, and also judge whether the image contains sea and land regions or only sea regions。 If only sea regions exist in image, we first use Gaussian low-pass filter to reduce stripe noises influences, then process the connected regions to detect ships。 If it is a sea-land image, we first do ship detections near the coast area and the sea area with the processing of connected regions, respectively; then we can obtain the final ship detection results by summing up the two parts’ results。
Keywords infrared remote sensing imagery, ship detection, region growing, morphological, connected region, stripe noise
本科毕业设计说明书 第 I 页
目 次
1 引言 1
1。1 选题背景及意义 1
1。3 本文主要内容及组织结构 4
2 图像预处理 5
2。1 图像地物为全海洋情况 5
2。2 图像地物为全海洋情况下预处理方法的选择 9
2。3 图像地物为海陆共存情况 11
2。4 图像地物为海陆共存情况下预处理方法的选择 12
3 图像分类和海陆分离 17
3。1 OTSU 算法 17
3。2