【摘要】 目标提取是对遥感图像进行分析处理的一个重要内容。针对彩色遥感图像的特征,首先采用了传统的边缘检测方法,验证其存在的不足。进而运用模糊聚类分析的模糊C-均值算法,选择适当的正交彩色特征,实现了对图像中目标区域的有效分割与快速提取,为遥感图像识别与计算提供了可靠的基础。研究实例表明了算法的有效性。更多还原

【Abstract】 As a comprehensive technology, remote sensing is more and more widely used in the constructing fields of national economy and defence. Especially in the applications of martial information scouting, the enemy objects information of posture, orientation, geometrical sizes and physical characteristics can be got by processing, recognizing and computing of the color sensing images from them. The information supplies the important basis for fight command, enemy object recognition and strike result evaluation. It is a complicated course to analyze and process the color remote sensing images and object extraction is an important part of it. Aiming at the characters of color remote sensing images, traditional methods of boundary detection, for example the arithmetic of Laplace, Sobel, Hough and so on, are applied for the typical example of remote sensing images to process at first in this paper and their deficiencies in use are illustrated by the examples. Then, Author adopts the fuzzy cluster method (FCM) and selects the appropriate color characters to segment the object areas from the images effectively and quickly. The investigation provides the reliable foundation for remote sensing images recognition and computation. Finally, the effectiveness of the method is demonstrated by the examples. The investigation of the paper was partially used in the strike result evaluation model of some type of weapon, and the result was satisfied with the requirement.更多还原