飞行器降落图像智能优化识别建模研究  

Aircraft Landing Image Intelligent Optimization Identification Modeling Research

作　　者：黄淼[1] 王刘涛[1] 褚龙现[1] 张海朝[2] 

机构地区：[1]平顶山学院软件学院,河南平顶山467000 [2]河南科技大学电子信息工程学院,河南洛阳471003

出　　处：《计算机仿真》2016年第1期86-89,共4页Computer Simulation

摘　　要：飞行器在复杂地形环境下的降落过程,需要根据采集的图像进行判断是否符合降落要求。飞行器最佳降落地点选择时,一旦降落地点复杂程度较高,采集的图像中将会产生海量的训练样本的数目,图像配准过程中,极大增加运算量,造成最佳降落地点选择的准确率降低。提出一种图像智能识别的最佳降落地点的选择方法。利用伪Zemike矩能够对降落地点的形状图像进行准确描述,利用Procrustes形状分析法提取最佳降落地点的特征,利用Rank的融合决策法最终完成最佳降落地点的选择。实验结果表明,利用改进算法进行飞行器最佳降落地点的选择,能够提高各地形最佳降落地点的准确率。Aircraft landing process in complex terrain environment, need according to the collected images to determine whether conform to the requirements of the landing. Aircraft landing site best choice, once landed complexity is higher, collection of images will produce large amounts of the number of training samples, the process of image reg- istration, greatly increase the computation, which reduces the best landing site selection accuracy. Put forward an in- telligent image recognition method to select the best landing sites. Using pseudo Zemike moment can carry on the ac- curate description to the landing place, the shape of the image, the characteristics of the shape of Procrustes analysis method is utilized to extract the best landing site, using the Rank of fusion decision final best landing site selection. Experimental results show that the improved algorithm for aircraft, the selection of optimal landing site can improve the accuracy of each terrain best landing site.

关 键 词：图像处理 智能识别 降落地点 

分 类 号：P631[天文地球—地质矿产勘探]