摘  要faint
遥感科学是随着现代科学技术的进步而发展起来的一门新兴学科，现已广泛应用于农业、林业、水利、气象、测绘、地质勘探、环境监测以及军事侦查等领域，并取得了很好的经济效益和社会效益。其中，从可见光遥感图像中自动获取有用信息一直是军事侦查领域的重要课题。在高技术条件下的局部战争中，飞机发挥着十分重要的作用，高效快速的飞机目标识别算法有利于作战指挥员实时把握敌方动态、进行分析决策进而赢得战争的胜利。因此，飞机目标的识别研究在军事应用上意义重大。
本文研究了可见光图像中地面飞机目标的识别算法。其中包括可见光图像的预处理和目标图像的区域分割，目标区域的特征选取和提取以及目标识别三个方面。在可见光图像的预处理和目标图像的区域分割方面，本文提出了基于模糊差影法的自适应阈值分割算法进行图像分割。在特征选取和提取算法的研究方面，提出了目标区域所占面积比这一特征量，并采用该特征量和紧凑度、七阶Hu不变矩对分割后的飞机目标区域进行了分析，消除了目标位置、方向、大小等对特征提取的干扰。最后，利用模糊模式识别算法进行飞机目标的识别，并标记出目标飞机。用本文提出的算法对搜集到的100幅图像进行处理，实验结果表明，识别率可达85%以上。
本文提出了基于模糊差影法的自适应阈值分割算法，并在识别过程中提出了目标区域所占面积比这一特征量，实现了可见光遥感图像中飞机目标的自动识别，并取得了较好的效果。
关键词：可见光图像；图像分割；目标特征提取；模糊模式识别
Aircraft Onground Recognition From Visible Images
Abstract
Remote sensing is a newly academic discipline which develops with modern science and technology. Now it is widely applied to the fields of agriculture, forestry, water conservancy, weather, measurement and mapping, the geology exploration, environment monitor and military investigation and so on, and it has obtained good economic benefits and social efficiency. As is known to all, it is important to obtain useful information automatically from visual images in the field of military investigation. In the local war under the high technique conditions, aircraft is of great importance. The efficient and high
rapid algorithms of airplane target recognition have a great advantage for chief officers to obtain information from enemy and thus win the war. Therefore, the research of aircraft target recognition is of great significance in military affairs.
In this paper, methods of aircraft recognition of visible images are discussed. Image preprocessed and image segment, target characteristic extraction and airplane target recognition are included. In the aspect of preprocessed and image segment, faintness difference and adaptive threshold value means are used to distill target section. In the aspect of target characteristic extraction, Hu’s seventh order invariant moment、 compactness and square rate are used to analyze the target section after segment. At last, we use fuzzy pattern recognition to recognize the airplane.
In this paper, faintness difference and adaptive threshold value means and target square rate are put forward and aircrafts in visible images can be recognition automatically.
.
Key Words：visible images ;  image segment ;  characteristic extraction ;  fuzzy pattern recognition

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。