Divided remote sensing image parallel processing method under Hadoop

一种Hadoop下的剖分遥感影像并行处理方法

Abstract

本发明属于影像处理技术领域,公开了一种Hadoop下的剖分遥感影像并行处理方法,包括:部署无线传感器,选择簇头,分簇,簇内节点构成简单图模型:通过S103得到簇内所有节点在簇内所处的位置,将每个节点当做图的一个顶点,每两个相邻节点间用边相连接;所述简单图模型为图像的显著性模型;簇内权值的计算;簇内权值的计算后进行网络数据认证;网络数据认证后进行数据库的缓存管理。本发明解决了传统的SVM分类方法的构造基于单机环境,受限于CPU的计算能力和内存的大小,只能串行处理遥感影像文件,仍然需要消耗很长的时间的问题;本发明图像数据处理效果清晰,准确率高。
The invention belongs to the technical field of image processing and discloses a divided remote sensing image parallel processing method under Hadoop. The method comprises the steps of deploying wireless sensors; selecting cluster heads; carrying out clustering; forming a simple map model by intra-cluster nodes, wherein the positions of all intra-cluster nodes in the clusters are obtained through S103, each node is taken as a vertex of a map, every two adjacent nodes are connected by a side, and the sample map model is a salient model of an image; calculating intra-cluster weights; carrying out network data verification after the intra-cluster weights are calculated; and carrying out cache management on a database after the network data verification is carried out. According to the method, the problem that a traditional SVM (Support Vector Machine) classification method is constructed based on a standalone environment, under the limitation of the computing capability of a CPU and a size of a memory, remote sensing image files only can be processed in series and very long time needs to be consumed is solved. The method is clear in image data processing effect and high in accuracy.

Claims

Description

Topics

Download Full PDF Version (Non-Commercial Use)

Patent Citations (0)

    Publication numberPublication dateAssigneeTitle

NO-Patent Citations (0)

    Title

Cited By (0)

    Publication numberPublication dateAssigneeTitle