Target recognition of aircraft based on moment invariants and BP neural network
Target recognition of aircraft based on moment invariants and BP neural network
复制标题
DOI：
--
复制
发表时间：
2012-06
期刊：
影响因子：
--
通讯作者：
Zixuan Fang;Guoqing Yao;Yanan Zhang
中科院分区：
文献类型：
--
作者：
Zixuan Fang;Guoqing Yao;Yanan Zhang
计算机视觉试图建立一种从图像或多维数据中获取信息的人工智能系统，目标识别是计算机视觉的一个重要研究领域。本文研究了如何在经过一系列预处理后实现对遥感图像中可疑目标区域的识别。采用轮廓跟踪法去除图像中的干扰区域，留下悬浮目标区域；从样本中提取飞机的矩不变量进行归一化，然后训练BP神经网络用于识别遥感图像中的可疑区域。本实验得到了符合识别要求的有效识别结果。
Computer vision is trying to build up an artificial intelligence system to obtain information from image or multidimensional data, target recognition of Computer Vision is one of an important research area. This paper studies how to realize recognition the suspected target area in remote sensing image after a series of pre-processing. Remove the interference area and leave suspended target area in the image by contour tracing method; normalizing moment invariants of aircraft by extracting them from sample, then training the BP neural network for the purpose of recognizing the suspected area in remote sensing image. We have got available recognition results which comply with the identification requirements with this experiment.